Brian Tinnell Keller
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Brian Tinnell Keller has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
- Craig K. Enders
Department of Psychology, University of California, Los Angeles
4 shared publications - Han Du
Department of Psychology, University of California, Los Angeles
2 shared publications - Jihyun Lee
University of North Texas
2 shared publications - S. Natasha Beretvas
University of Texas at Austin
2 shared publications - Benjamin R. Karney
Department of Psychology, University of California, Los Angeles
1 shared publication Craig K. Enders
Department of Psychology, University of California, Los Angeles
1 shared publication
Latest Publications
- Open Access
- 8 September 2026
- Published by Center for Open Science in PsyArXiv
Dynamic Structural Equation Modeling in Blimp
Dynamic structural equation modeling (DSEM) provides a framework for studying temporal processes and individual differences in intensive longitudinal data. This article presents a tutorial on fitting and interpreting these models with Blimp, a freely available statistical program for all major platforms. Blimp’s factored-regression formulation expresses dynamic models as linked conditional regressions, connecting familiar multilevel notation to flexible specifications involving latent person means, dynamic coefficients, and residual variances. These person-specific quantities can enter other equations as predictors or moderators, allowing researchers to investigate relationships among typical levels, temporal dynamics, and variability within a common framework. Residual variance equations additionally accommodate both within-person and between-person predictors, making variability itself an outcome to be explained. The article also extends familiar conditional-effect calculations and graphics to latent interactions and random-slope heterogeneity, providing tools for interpreting individual differences in dynamic relationships that apply beyond DSEM. Using daily diary data on sleep quality and chronic pain, the tutorial connects these capabilities to substantive research questions, model equations, software syntax, and interpretation. Complete data, Blimp and R scripts, output, and supplemental explanations provide resources for applying and extending the framework.
- 8 September 2026
- Published by Center for Open Science in PsyArXiv
Open AccessDynamic Structural Equation Modeling in Blimp
Dynamic structural equation modeling (DSEM) provides a framework for studying temporal processes and individual differences in intensive longitudinal data. This article presents a tutorial on fitting and interpreting these models with Blimp, a freely available statistical program for all major platforms. Blimp’s factored-regression formulation expresses dynamic models as linked conditional regressions, connecting familiar multilevel notation to flexible specifications involving latent person means, dynamic coefficients, and residual variances. These person-specific quantities can enter other equations as predictors or moderators, allowing researchers to investigate relationships among typical levels, temporal dynamics, and variability within a common framework. Residual variance equations additionally accommodate both within-person and between-person predictors, making variability itself an outcome to be explained. The article also extends familiar conditional-effect calculations and graphics to latent interactions and random-slope heterogeneity, providing tools for interpreting individual differences in dynamic relationships that apply beyond DSEM. Using daily diary data on sleep quality and chronic pain, the tutorial connects these capabilities to substantive research questions, model equations, software syntax, and interpretation. Complete data, Blimp and R scripts, output, and supplemental explanations provide resources for applying and extending the framework.
- 1 August 2026
- Published by American Psychological Association (APA) in Psychological Methods
- Vol. 31 (4) , 724-748
Factored structural equation modeling in blimp.
This tutorial introduces factored structural equation modeling (FSEM), an alternative to multivariate structural equation modeling, and demonstrates its implementation in the Blimp software and the rblimp R package. FSEM reconceptualizes the joint distribution of observed and latent variables as a set of univariate and multivariate submodels-each specified via simple regression equations-and treats latent variables as missing data to be imputed via Bayesian data augmentation. This approach seamlessly accommodates combinations of continuous (normal and nonnormal), binary, ordinal, nominal, count, and two-part variables; interactions; nonlinear effects; heteroscedasticity; and multilevel data structures without violating distributional assumptions. We first outline the theoretical foundations of factored regression specification and its connection to the probability chain rule. We then describe the Markov chain Monte Carlo estimation and missing-data imputation algorithms that underlie FSEM. Next, we present a series of illustrative models-ranging from basic confirmatory factor analysis to complex dynamic, multilevel, and hybrid generalized-linear-structural equation modeling applications-providing Blimp syntax excerpts and a real-data example. We conclude by discussing practical considerations, limitations, and directions for future methodological research. This tutorial aims to equip researchers with a flexible, user-friendly framework for modeling complex data in behavioral and social sciences. (PsycInfo Database Record (c) 2026 APA, all rights reserved).
- 1 August 2026
- Published by American Psychological Association (APA) in Psychological Methods
- Vol. 31 (4) , 724-748
Factored structural equation modeling in blimp.
This tutorial introduces factored structural equation modeling (FSEM), an alternative to multivariate structural equation modeling, and demonstrates its implementation in the Blimp software and the rblimp R package. FSEM reconceptualizes the joint distribution of observed and latent variables as a set of univariate and multivariate submodels-each specified via simple regression equations-and treats latent variables as missing data to be imputed via Bayesian data augmentation. This approach seamlessly accommodates combinations of continuous (normal and nonnormal), binary, ordinal, nominal, count, and two-part variables; interactions; nonlinear effects; heteroscedasticity; and multilevel data structures without violating distributional assumptions. We first outline the theoretical foundations of factored regression specification and its connection to the probability chain rule. We then describe the Markov chain Monte Carlo estimation and missing-data imputation algorithms that underlie FSEM. Next, we present a series of illustrative models-ranging from basic confirmatory factor analysis to complex dynamic, multilevel, and hybrid generalized-linear-structural equation modeling applications-providing Blimp syntax excerpts and a real-data example. We conclude by discussing practical considerations, limitations, and directions for future methodological research. This tutorial aims to equip researchers with a flexible, user-friendly framework for modeling complex data in behavioral and social sciences. (PsycInfo Database Record (c) 2026 APA, all rights reserved).
- 26 May 2026
- Published by Taylor & Francis in Multivariate Behavioral Research
- Vol. 61 (5) , 1-25
Exploring the Use of Multiple Imputation for Handling Missing Covariates in Meta-Regression with Dependent Effect Sizes
Meta-analysis synthesizes empirical evidence from multiple studies that examine comparable research topics to provide a summary and explore heterogeneity in the findings of the studies. Meta-analyt...
- 26 May 2026
- Published by Taylor & Francis in Multivariate Behavioral Research
- Vol. 61 (5) , 1-25
Exploring the Use of Multiple Imputation for Handling Missing Covariates in Meta-Regression with Dependent Effect Sizes
Meta-analysis synthesizes empirical evidence from multiple studies that examine comparable research topics to provide a summary and explore heterogeneity in the findings of the studies. Meta-analyt...
- Open Access
- 22 May 2026
- Published by Center for Open Science in PsyArXiv
Exploring the Use of Multiple Imputation for Handling Missing Covariates in Meta-regression with Dependent Effect Sizes
Meta-analysts frequently encounter missing covariate values, which can complicate valid estimation of meta-regression models. In practice, missing data are managed often through ad hoc deletion approaches, which can reduce the validity of statistical inferences. More advanced missing data handling approaches such as multiple imputation (MI) remain underutilized, particularly in meta-analyses with dependent effect sizes within studies. This study expands the use of MI techniques for handling missing covariates in such contexts. Specifically, this study introduces adapted multilevel MI approaches that are expected to better accommodate the structure of meta-analytic data with dependent effect sizes and associated model. The study presents Monte Carlo computer simulations that compare the performance of different MI techniques including single-level and multilevel agnostic MI methods as well as multilevel substantive model-based MI and ad hoc deletion approaches. The results generally supported the use of the MI approach over deletion approaches when the dependent structure is well specified in the imputation procedure. This study demonstrates the feasibility of the use of MI techniques and underscores the importance of incorporating the hierarchical structure into the meta-regression model when analyzing dependent effect sizes with missing covariate values. Implications and directions for future research are discussed.
- 22 May 2026
- Published by Center for Open Science in PsyArXiv
Open AccessExploring the Use of Multiple Imputation for Handling Missing Covariates in Meta-regression with Dependent Effect Sizes
Meta-analysts frequently encounter missing covariate values, which can complicate valid estimation of meta-regression models. In practice, missing data are managed often through ad hoc deletion approaches, which can reduce the validity of statistical inferences. More advanced missing data handling approaches such as multiple imputation (MI) remain underutilized, particularly in meta-analyses with dependent effect sizes within studies. This study expands the use of MI techniques for handling missing covariates in such contexts. Specifically, this study introduces adapted multilevel MI approaches that are expected to better accommodate the structure of meta-analytic data with dependent effect sizes and associated model. The study presents Monte Carlo computer simulations that compare the performance of different MI techniques including single-level and multilevel agnostic MI methods as well as multilevel substantive model-based MI and ad hoc deletion approaches. The results generally supported the use of the MI approach over deletion approaches when the dependent structure is well specified in the imputation procedure. This study demonstrates the feasibility of the use of MI techniques and underscores the importance of incorporating the hierarchical structure into the meta-regression model when analyzing dependent effect sizes with missing covariate values. Implications and directions for future research are discussed.
- 30 October 2025
- Published by American Psychological Association (APA) in Psychological Methods
A factored regression approach to modeling latent variable interactions and nonlinear effects.
Interaction effects are common in the behavioral sciences, especially in psychology, as they help explore how various factors influence human behavior. This article introduces a factored regression framework designed to estimate latent variable interactions and nonlinear effects, providing a flexible approach for modeling complex data structures, accommodating diverse data types, and handling missing data on any variable. The factored regression framework also allows graphical diagnostics to probe interactions effectively. Monte Carlo simulations were conducted to compare the performance of factored regression with existing maximum likelihood methods, such as latent moderated structural equations and product indicators. Results indicate that factored regression performs comparably to, if not better than, these traditional methods. The factored regression framework is implemented in Blimp software, offering an accessible and user-friendly syntax for specifying the models. Through practical examples and syntax excerpts, this article demonstrates the application of factored regression for estimating latent interactions, making it more approachable for a wide audience in behavioral research. (PsycInfo Database Record (c) 2026 APA, all rights reserved).
- 30 October 2025
- Published by American Psychological Association (APA) in Psychological Methods
A factored regression approach to modeling latent variable interactions and nonlinear effects.
Interaction effects are common in the behavioral sciences, especially in psychology, as they help explore how various factors influence human behavior. This article introduces a factored regression framework designed to estimate latent variable interactions and nonlinear effects, providing a flexible approach for modeling complex data structures, accommodating diverse data types, and handling missing data on any variable. The factored regression framework also allows graphical diagnostics to probe interactions effectively. Monte Carlo simulations were conducted to compare the performance of factored regression with existing maximum likelihood methods, such as latent moderated structural equations and product indicators. Results indicate that factored regression performs comparably to, if not better than, these traditional methods. The factored regression framework is implemented in Blimp software, offering an accessible and user-friendly syntax for specifying the models. Through practical examples and syntax excerpts, this article demonstrates the application of factored regression for estimating latent interactions, making it more approachable for a wide audience in behavioral research. (PsycInfo Database Record (c) 2026 APA, all rights reserved).
- 1 October 2025
- Published by American Psychological Association (APA) in Psychological Methods
- Vol. 30 (5) , 980-996
A simple Monte Carlo method for estimating power in multilevel designs.
Estimating power for multilevel models is complex because there are many moving parts, several sources of variation to consider, and unique sample sizes at Level 1 and Level 2. Monte Carlo computer simulation is a flexible tool that has received considerable attention in the literature. However, much of the work to date has focused on very simple models with one predictor at each level and one cross-level interaction effect, and approaches that do not share this limitation require users to specify a large set of population parameters. The goal of this tutorial is to describe a flexible Monte Carlo approach that accommodates a broad class of multilevel regression models with continuous outcomes. Our tutorial makes three important contributions. First, it allows any number of within-cluster effects, between-cluster effects, covariate effects at either level, cross-level interactions, and random coefficients. Moreover, we do not assume orthogonal effects, and predictors can correlate at either level. Second, our approach accommodates models with multiple interaction effects, and it does so with exact expressions for the variances and covariances of product random variables. Finally, our strategy for deriving hypothetical population parameters does not require pilot or comparable data. Instead, we use intuitive variance-explained effect size expressions to reverse-engineer solutions for the regression coefficients and variance components. We describe a new R package mlmpower that computes these solutions and automates the process of generating artificial data sets and summarizing the simulation results. The online supplemental materials provide detailed vignettes that annotate the R scripts and resulting output. (PsycInfo Database Record (c) 2023 APA, all rights reserved).
- 1 October 2025
- Published by American Psychological Association (APA) in Psychological Methods
- Vol. 30 (5) , 980-996
A simple Monte Carlo method for estimating power in multilevel designs.
Estimating power for multilevel models is complex because there are many moving parts, several sources of variation to consider, and unique sample sizes at Level 1 and Level 2. Monte Carlo computer simulation is a flexible tool that has received considerable attention in the literature. However, much of the work to date has focused on very simple models with one predictor at each level and one cross-level interaction effect, and approaches that do not share this limitation require users to specify a large set of population parameters. The goal of this tutorial is to describe a flexible Monte Carlo approach that accommodates a broad class of multilevel regression models with continuous outcomes. Our tutorial makes three important contributions. First, it allows any number of within-cluster effects, between-cluster effects, covariate effects at either level, cross-level interactions, and random coefficients. Moreover, we do not assume orthogonal effects, and predictors can correlate at either level. Second, our approach accommodates models with multiple interaction effects, and it does so with exact expressions for the variances and covariances of product random variables. Finally, our strategy for deriving hypothetical population parameters does not require pilot or comparable data. Instead, we use intuitive variance-explained effect size expressions to reverse-engineer solutions for the regression coefficients and variance components. We describe a new R package mlmpower that computes these solutions and automates the process of generating artificial data sets and summarizing the simulation results. The online supplemental materials provide detailed vignettes that annotate the R scripts and resulting output. (PsycInfo Database Record (c) 2023 APA, all rights reserved).
- Open Access
- 28 June 2025
- Published by Center for Open Science in PsyArXiv
Factored Structural Equation Modeling in Blimp
This tutorial introduces Factored Structural Equation Modeling (FSEM), an alternative to classic multivariate SEM, and demonstrates its implementation in the Blimp software and the rblimp R package. FSEM reconceptualizes the joint distribution of observed and latent variables as a set of univariate and multivariate submodels—each specified via simple regression equations—and treats latent variables as missing data to be imputed via Bayesian data augmentation. This approach seamlessly accommodates combinations of continuous (normal and nonnormal), binary, ordinal, nominal, count, and two-part variables; interactions; nonlinear effects; heteroscedasticity; and multilevel data structures without violating distributional assumptions. We first outline the theoretical foundations of factored regression specification and its connection to the probability chain rule. We then describe the Markov chain Monte Carlo estimation and missing-data imputation algorithms that underlie FSEM. Next, we present a series of illustrative models—ranging from basic confirmatory factor analysis to complex dynamic, multilevel, and hybrid generalized-linear-SEM applications—providing Blimp syntax excerpts and a real-data example. We conclude by discussing practical considerations, limitations, and directions for future methodological research. This tutorial aims to equip researchers with a flexible, user-friendly framework for modeling complex data in behavioral and social sciences.
- 28 June 2025
- Published by Center for Open Science in PsyArXiv
Open AccessFactored Structural Equation Modeling in Blimp
This tutorial introduces Factored Structural Equation Modeling (FSEM), an alternative to classic multivariate SEM, and demonstrates its implementation in the Blimp software and the rblimp R package. FSEM reconceptualizes the joint distribution of observed and latent variables as a set of univariate and multivariate submodels—each specified via simple regression equations—and treats latent variables as missing data to be imputed via Bayesian data augmentation. This approach seamlessly accommodates combinations of continuous (normal and nonnormal), binary, ordinal, nominal, count, and two-part variables; interactions; nonlinear effects; heteroscedasticity; and multilevel data structures without violating distributional assumptions. We first outline the theoretical foundations of factored regression specification and its connection to the probability chain rule. We then describe the Markov chain Monte Carlo estimation and missing-data imputation algorithms that underlie FSEM. Next, we present a series of illustrative models—ranging from basic confirmatory factor analysis to complex dynamic, multilevel, and hybrid generalized-linear-SEM applications—providing Blimp syntax excerpts and a real-data example. We conclude by discussing practical considerations, limitations, and directions for future methodological research. This tutorial aims to equip researchers with a flexible, user-friendly framework for modeling complex data in behavioral and social sciences.
- Open Access
- 11 June 2025
- Published by Center for Open Science in PsyArXiv
A Factored Regression Approach to Modeling Latent Variable Interactions and Nonlinear Effects
Interaction effects are common in the behavioral sciences, especially in psychology, as they help explore how various factors influence human behavior. This paper introduces a factored regression framework designed to estimate latent variable interactions and nonlinear effects, providing a flexible approach for modeling complex data structures, accommodating diverse data types, and handling missing data on any variable. The factored regression framework also allows graphical diagnostics to probe interactions effectively. Monte Carlo simulations were conducted to compare the performance of factored regression with existing maximum likelihood methods, such as latent moderated structural equations and product indicators. Results indicate that factored regression performs comparably to, if not better than, these traditional methods. The factored regression framework is implemented in Blimp software, offering an accessible and user-friendly syntax for specifying the models. Through practical examples and syntax excerpts, this article demonstrates the application of factored regression for estimating latent interactions, making it more approachable for a wide audience in behavioral research.
- 11 June 2025
- Published by Center for Open Science in PsyArXiv
Open AccessA Factored Regression Approach to Modeling Latent Variable Interactions and Nonlinear Effects
Interaction effects are common in the behavioral sciences, especially in psychology, as they help explore how various factors influence human behavior. This paper introduces a factored regression framework designed to estimate latent variable interactions and nonlinear effects, providing a flexible approach for modeling complex data structures, accommodating diverse data types, and handling missing data on any variable. The factored regression framework also allows graphical diagnostics to probe interactions effectively. Monte Carlo simulations were conducted to compare the performance of factored regression with existing maximum likelihood methods, such as latent moderated structural equations and product indicators. Results indicate that factored regression performs comparably to, if not better than, these traditional methods. The factored regression framework is implemented in Blimp software, offering an accessible and user-friendly syntax for specifying the models. Through practical examples and syntax excerpts, this article demonstrates the application of factored regression for estimating latent interactions, making it more approachable for a wide audience in behavioral research.
- Open Access
- 13 December 2024
- Published by Center for Open Science in PsyArXiv
Blimp Technical Appendix: Nonnormal Latent Variables using the Yeo-Johnson Transformation
This technical appendix describes modeling asymmetrical nonnormal latent variables. Such asymmetrically distributed latent traits can map onto many researchers’ intuitions about the underlying beliefs about the trait in the population. Examples in the psychological literature include dark triad traits (Webster & Jonason, 2013), borderline personality disorder (Michonski, Sharp, Steinberg, & Zanarini, 2013), and self-esteem (Gray-Little, Williams, & Hancock, 1997), among others. Blimp allows for the modeling of asymmetrical latent traits via transformation methods. While any invertible transformation can be implemented, this technical appendix aims to describe the application of the Yeo-Johnson transformation (Yeo & Johnson, 2000).
- 13 December 2024
- Published by Center for Open Science in PsyArXiv
Open AccessBlimp Technical Appendix: Nonnormal Latent Variables using the Yeo-Johnson Transformation
This technical appendix describes modeling asymmetrical nonnormal latent variables. Such asymmetrically distributed latent traits can map onto many researchers’ intuitions about the underlying beliefs about the trait in the population. Examples in the psychological literature include dark triad traits (Webster & Jonason, 2013), borderline personality disorder (Michonski, Sharp, Steinberg, & Zanarini, 2013), and self-esteem (Gray-Little, Williams, & Hancock, 1997), among others. Blimp allows for the modeling of asymmetrical latent traits via transformation methods. While any invertible transformation can be implemented, this technical appendix aims to describe the application of the Yeo-Johnson transformation (Yeo & Johnson, 2000).
- Open Access
- 8 November 2024
- Published by Center for Open Science in PsyArXiv
mlmpower: An R Package for Conducting Power Analysis and Data Simulation with Multilevel Models
Power analysis is a critical step in designing studies involving multilevel models, yet existing approaches can be complex and limited in flexibility. This paper introduces mlmpower, an R package designed to simplify Monte Carlo power analysis for multilevel models. The package accommodates various model specifications, including cross-level interactions, random slopes, and within-cluster and between-cluster predictors. Unlike other methods, which often require a pilot study or extensive prior data, mlmpower uses intuitive effect size metrics to simulate data and estimate power with confidence intervals. Inspired by ggplot2, the package's syntax allows for modular model building, making it accessible even to researchers with minimal simulation experience. Additionally, mlmpower supports including missing data mechanisms, offering more realistic power estimates for longitudinal and cross-sectional studies.
- 8 November 2024
- Published by Center for Open Science in PsyArXiv
Open Accessmlmpower: An R Package for Conducting Power Analysis and Data Simulation with Multilevel Models
Power analysis is a critical step in designing studies involving multilevel models, yet existing approaches can be complex and limited in flexibility. This paper introduces mlmpower, an R package designed to simplify Monte Carlo power analysis for multilevel models. The package accommodates various model specifications, including cross-level interactions, random slopes, and within-cluster and between-cluster predictors. Unlike other methods, which often require a pilot study or extensive prior data, mlmpower uses intuitive effect size metrics to simulate data and estimate power with confidence intervals. Inspired by ggplot2, the package's syntax allows for modular model building, making it accessible even to researchers with minimal simulation experience. Additionally, mlmpower supports including missing data mechanisms, offering more realistic power estimates for longitudinal and cross-sectional studies.