Latent Transition Analysis
Also known as: LTA
Latent Transition Analysis (LTA) is a method for studying transitions between latent classes over time, developed by Collins and Lanza (2010). LTA combines latent class analysis (grouping individuals into classes) with Markovian transition models to understand how people move between qualitatively distinct states across time periods.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
+1 more
When to use it
Apply LTA when studying developmental processes with discrete stages, tracking changes in qualitative group membership over time, or predicting transitions between states. Ideal for longitudinal studies with repeated measurements over multiple time points.
Strengths & limitations
- Captures transitions: reveals movement between qualitative states, not just average change
- Discrete states: identifies distinct classes rather than assuming continuous variation
- Probabilistic transitions: allows uncertainty in transitions and heterogeneous patterns
- Longitudinal: naturally addresses within-person change over time
- Predictive: can identify covariates predicting transitions or stability
- Latent class assumptions: assumes discrete classes exist; may not match continuous reality
- Markovian assumption: assumes future depends only on present, not full history (often violated)
- Number of classes: choosing number of latent classes is subjective
- Sample size: requires substantial longitudinal data (multiple time points, large n)
Frequently asked
What is the difference between latent class analysis (LCA) and latent transition analysis (LTA)?
LCA identifies classes at a single time point. LTA extends this to multiple time points and models transitions between classes over time. LTA is the longitudinal extension of LCA.
What is a Markovian transition assumption?
It assumes the probability of being in class B at time t+1 depends only on class membership at time t, not on the full history. This is often unrealistic; violations reduce model validity.
How do I choose the number of latent classes?
Use fit indices (BIC, AIC) and interpretability. Start with 2 classes and increase until fit indices stop improving or classes become uninterpretable. Validate with external criteria.
Can LTA include covariates?
Yes. Covariates can predict baseline class membership or transition probabilities. This allows examining who is at risk for certain transitions.
What if someone has missing data at one time point?
LTA can handle missing data under missing-at-random assumption using maximum likelihood. Multiple imputation is another option. Avoid listwise deletion as it biases results.
Sources
- Collins, L. M., & Lanza, S. T. (2010). Latent Class and Latent Transition Analysis: With Applications in the Social, Behavioral, and Health Sciences. Wiley. ISBN: 9780470228395
- Lanza, S. T., Collins, L. M., Lemmon, D. R., & Schafer, J. L. (2007). PROC LTA: A SAS macro for latent transition analysis. Structural Equation Modeling, 14(4), 671-694. link ↗
- Vermunt, J. K., & Magidson, J. (2016). Latent class and latent transition analysis. In J. P. Baltes, G. G. Brim, D. Featherman, & S. Shye (Eds.), Lifespan Development and Behavior (pp. 91-113). Academic Press. ISBN: 9780123997760
How to cite this page
ScholarGate. (2026, June 3). Latent Transition Analysis. ScholarGate. https://scholargate.app/en/psychometrics/latent-transition-analysis
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- DINA ModelPsychometrics↔ compare
- Exploratory Structural Equation ModelingPsychometrics↔ compare
- Partial Least Squares Structural Equation ModelingPsychometrics↔ compare
- Value-Added ModelingPsychometrics↔ compare