Introduction to Latent Profile Transition Analysis

A hands-on and non-technical introduction to the statistical technique of latent profile transition analysis – the longitudinal extension of latent profile analyses. Designed to help you understand why, when, and how to conduct LPTA!

Instructor:
Sara K. Johnson, PhD (Tufts University)

⚠️ Workshop Format: ⚠️
Three-Day Asynchronous Workshop with Office Hours

Workshop Videos Released:
Tuesday, July 1, 2025

Optional Office Hours/Discussion Meetings:
Monday, August 11, 2025 at 4:00pm-5:00pm EDT
Tuesday, August 12, 2025 at 3:00pm-4:00pm EDT
Wednesday, August 13, 2025 at 1:00pm-2:00pm EDT

Many researchers are interested in identifying groups of people who share similarities based on a set of beliefs, behaviors, or attitudes (e.g., parenting styles captured by warmth and monitoring). These types of research questions pose analysis challenges because the subgroups cannot be captured directly with a single question but instead must be inferred from multiple pieces of information (in other words, group membership is latent). Mixture models (a group of statistical analyses) can be used in these situations because they are designed to investigate latent subgroups. These models have become quite popular in the social and behavioral sciences but are less frequently included in statistical coursework compared to other techniques (e.g., multiple regression, factor analyses).

This workshop addresses the basics of conducting Latent Profile Transition Analysis (LPTA). It begins where the Introduction to Latent Profile Analysis (LPTA) workshop ends and covers how to link LPAs from multiple time points into a longitudinal analysis. Accordingly, the Introduction to Latent Profile Transition Analysis workshop is designed for attendees who are already familiar with LPA.

The LPTA workshop introduces important aspects of this analysis technique in a non-technical way: 1. conceptual foundations, including understanding when an LPTA is an appropriate choice (compared to other types of longitudinal analyses), and formulating suitable research questions; 2. data requirements, including planning for data collection or evaluating the suitability of previously collected data; 3. conducting analyses, with step-by-step practice; and 4. communicating results through text and visuals. The workshop also includes a small amount of introductory information about LPTA extensions, including adding predictors and outcomes of profile membership and transitions between profiles. The concepts and procedures of LPTA will be illustrated using a simulated data set about a popular young adult book series.

Registered learners will have access to the lecture recordings (approximately 15 hours of content) and associated materials (worksheets and self-test activities, downloadable slideshows, example data files, example syntax files, and more) as well as the passcode to attend the optional synchronous question and discussion sessions held on Zoom.

Syllabus

  • Conceptual Background

    • When a latent profile transition analysis (LPTA) approach is appropriate

    • The research questions that can be addressed with LPTA

  • Data Requirements

    • Sufficient sample sizes

    • The role of statistical power in LPTA

    • The number and type of items that are needed

  • Data Evaluation and Preparation

    • How missing data are dealt with in LPTA

    • The role of item distributions

    • The role of inter-item correlations

  • Conducting Analyses

    • Procedures for testing whether the number and type of profiles are the same across time points

    • How to accommodate similarities and differences across time points in the number and type of profiles

    • Efficient tracking of information from model testing

    • Decision-making processes for choosing the final LPTA model

  • Presenting Results

    • Visual and text-based options for presenting profile results

    • Visual options for presenting transitions between profiles

    • The kinds of information to include in a manuscript

    • The kinds of information to include in an appendix

  • Extensions of Latent Profile Transition Analysis

    • Adding predictors and outcomes of profile membership

    • Adding predictors of transition probabilities

    • Testing for differences in outcome variables based on transitions

    • Multiple-Group Models

Registration Options

Introduction to Latent Profile Transition Analysis

  • Professional
  • $674
  • Baseline Price for Faculty,
    Staff, and Other Professionals
  • Click Register Below
  • Trainee
  • $674 $452
  • 33% Discount for
    Students and Postdocs
  • Use code "TRAINEE" at Checkout

Combo 1: Introduction to Latent Profile Transition Analysis + Introduction to Latent Profile Analysis

  • Professional
  • $1348 $1078
  • Baseline Price for Faculty,
    Staff, and Other Professionals
  • 20% Combination Discount
  • Click Register Below
  • Trainee
  • $1078 $723
  • 33% Discount for
    Students and Postdocs
  • 20% Combination Discount
  • Use code "TRAINEE" at Checkout

 FAQs

  • This workshop is designed for researchers in the social, behavioral, and medical sciences who have some experience with statistical analysis, including understanding Latent Profile Analysis. However, no background with longitudinal analyses more generally is required. The material focuses on conceptual and pragmatic aspects of LPTA rather than statistical technicalities. Accordingly, this workshop is appropriate for faculty, graduate students, and advanced undergraduate students with sufficient prior knowledge.

  • Intermediate

  • This workshop assumes that attendees have completed the Introduction to Latent Profile Analysis Workshop (a combination discount is available to learners who register for both) or already know all the material taught within it (i.e., formulating research questions, evaluating data, and conducting analyses. It is especially important that attendees understand the importance of, and procedures for, testing various forms of the within-profile variance-covariance matrices).

  • Analyses will be demonstrated in Mplus, and input and output files will be available for Mplus as well. A subset of the models discussed (but not all of them) can be estimated in other software (such as Latent Gold, SAS, or R). Accordingly, I unfortunately can't answer questions about how to estimate these models in software other than Mplus. I suggest that if you want to do that, you start with trying to replicate the same models we discussed in the workshop (using the same data set) so that you can compare the results and will know what to look for when estimating models with your own data.

  • Registered learners will have access to the lecture recordings (approximately 15 hours of content) and associated materials (worksheets and self-test activities, downloadable slideshows, example data files, example syntax files, and more) as well as the passcode to attend the optional synchronous question and discussion sessions held on Zoom.