Methodological & Technical Research Topics | The Methodology Center

Methodological & Technical Research Topics

LCA With a Distal Outcome

The %LCA_Distal macro provides a model-based approach to estimating the association between C and Z.

Latent class membership can be used to predict a distal outcome (an outcome at a later time). The Methodology Center has developed a comprehensive, model-based approach to estimating the effect of latent class membership on a distal outcome (Lanza, Tan, & Bray, 2013). This approach can be implemented using our %LCA_Distal SAS macro. We have also developed an approach to improve the posterior probabilities on which classify-analyze approaches (e.g., modal assignment) are based (Bray, Lanza, & Tan, 2014).


References & recommended reading

Bray, B. C., Lanza, S. T., & Tan, X. (2014). Eliminating bias in classify-analyze approaches for latent class analysis. Structural Equation Modeling: A Multidisciplinary Journal. Advance online publication. doi: 10.1080/10705511.2014.935265. 

Lanza, S. T. & Rhoades, B. L. (2013). Latent class analysis: An alternative perspective on subgroup analysis in prevention and treatment. Prevention Science, 14, 157-168. PMCID: PMC3173585

Lanza, S. T., Tan, X., & Bray, B. C. (2013). Latent class analysis with distal outcomes: A flexible model-based approach. Structural Equation Modeling: A Multidisciplinary Journal,20, 1-26.


Latent Class Moderation

Moderation analysis is typically conducted by incorporating a single variable (e.g., gender, baseline severity) as a moderator into a multiple regression model. By using LCA, researchers can identify subgroups of people exposed to a common set of factors, and who, therefore, may respond differently to intervention.


References & recommended reading

Cooper, B. R., & Lanza, S. T. (2014). Who benefits most from Head Start? Using latent class moderation to examine differential treatment effects. Child Development. Advance online publication. doi: 10.1111/cdev.12278

Lanza, S. T. & Rhoades, B. L. (2013). Latent class analysis: An alternative perspective on subgroup analysis in prevention and treatment. Prevention Science, 14, 157-168. PMCID: PMC3173585

Cleveland, M. J., Lanza, S. T., Ray, A. E., Turrisi, R., & Mallett, K. M. (2012). Transitions in first-year college student drinking behaviors: Does drinking latent class membership moderate the effects of parent- and peer-based intervention components? Psychology of Addictive Behaviors, 26, 440-450. PMCID: PMC3413757


LCA with Causal Inference

LCA with covariates models the association between classes and predictors, but causation cannot be inferred unless people are randomly assigned to levels of the predictor of latent class membership. Modern causal inference methods, such as inverse propensity weighting, can be used to adjust for potential confounding in observational data. The Methodology Center has pioneered work on applying inverse propensity weights to estimate the causal effects of covariates on latent class membership and to estimate the causal effects of latent class membership on a distal outcome. 


References & recommended reading

Schuler, M. S., Leoutsakos, J. S., & Stuart, E. A. (2014). Addressing confounding when estimating the effects of latent classes on a distal outcome. Health Services Outcomes and Research Methodology14(4), 232-254.

Lanza, S.T., Schuler, M.S., & Bray, B.C. (in press). Latent class analysis with causal inference: The effect of adolescent depression on young adult substance abuse profiles. In Causality and Statistics.

Butera, N. M., Lanza, S. T., & Coffman, D. L. (2013). A framework for estimating causal effects in latent class analysis: Is there a causal link between early sex and subsequent profiles of delinquency? Prevention Science. doi: 10.1007/s11121-013-0417-3  PMCID: PMC3888479

Lanza, S. T., Coffman, D. L., & Xu, S. (2013). Causal inference in latent class analysis. Structural Equation Modeling, 20(3), 361-383. PMCID: PMC4240500


Power in LCA

Applications of LCA require choosing the number of classes posited to exist in the population. When there is not an adequate theoretical basis for confidently specifying this number a priori, data-driven approaches are required for choosing the number of classes or testing whether a proposed number of classes is adequate. The bootstrap likelihood ratio test is considered one of the best and most rigorous approaches to making this decision.


References & recommended reading

Dziak, J. J., Lanza, S. T., & Tan, X. (2014). Effect size, statistical power and sample size requirements for the bootstrap likelihood ratio test in latent class analysis. Structural Equation Modeling, 21, 534-552. 


Software Development

We distribute free software to researchers so they can use LCA accurately and easily. The Methodology Center first released PROC LCA for SAS in 2007, and we have regularly added important features to the software.


We also develop macros to enhance PROC LCA functionality.


We also develop non-SAS software for LCA.

Like Us On Facebook or Tweet This Page