Wen Luo
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Wen Luo has a total of 50 co-authors. Here are the top co-authors by number of shared publications:
Eunkyeng Baek
Texas A&M University
4 shared publicationsChendong Li
James Madison University
3 shared publications- Chendong Li
Texas A&M University
3 shared publications Haoran Li
3 shared publications- Eduardo Gandara
Department of Public Health, UT San Antonio, San Antonio, USA
2 shared publications - Eunkyeng Baek
Department of Educational Psychology, Texas A&M University, College Station, USA
2 shared publications
Latest Publications
- Open Access
- 18 August 2026
- Published by Springer Nature in Research Square
Bridging SCED and Group Designs: Demonstrating the Ratio-Metric BC-IRR Effect Size in a Meta-Analysis
This study introduces the between-case incidence rate ratio (BC-IRR), a ratio-metric effect size for single-case experimental designs (SCEDs), and demonstrates its use in a reanalysis of Barton et al.’s (2017) meta-analysis of technology-aided instruction and intervention (TAII) for students with autism spectrum disorder (ASD). We applied generalized estimating equation (GEE) models to estimate BC-IRR for SCED outcomes and computed mean response ratios for group designs. Across 36 eligible outcomes (29 from five group-design studies and 8 from eight SCED studies), separate random-effects models showed that group-design studies yielded a pooled mean response ratio of 1.23 (95% CI [1.00, 1.52]), indicating approximately 23% higher positive outcomes under TAII relative to comparison conditions, whereas SCEDs produced a pooled BC-IRR of 4.52 (95% CI [1.71, 11.92]), indicating more than a fourfold increase over baseline. We discuss BC-IRR’s advantages over BC-SMD, namely, its invariance to outcome variance, insensitivity to observation session length, and direct percent-change interpretation, alongside practical limitations, including the requirement of nonzero baselines. These findings illustrate how BC-IRR can support design-comparable, interpretable synthesis of SCED and group-design evidence while underscoring the importance of triangulating ratio-metric and SMD-type effect sizes when evaluating intervention impact.
- 18 August 2026
- Published by Springer Nature in Research Square
Open AccessBridging SCED and Group Designs: Demonstrating the Ratio-Metric BC-IRR Effect Size in a Meta-Analysis
This study introduces the between-case incidence rate ratio (BC-IRR), a ratio-metric effect size for single-case experimental designs (SCEDs), and demonstrates its use in a reanalysis of Barton et al.’s (2017) meta-analysis of technology-aided instruction and intervention (TAII) for students with autism spectrum disorder (ASD). We applied generalized estimating equation (GEE) models to estimate BC-IRR for SCED outcomes and computed mean response ratios for group designs. Across 36 eligible outcomes (29 from five group-design studies and 8 from eight SCED studies), separate random-effects models showed that group-design studies yielded a pooled mean response ratio of 1.23 (95% CI [1.00, 1.52]), indicating approximately 23% higher positive outcomes under TAII relative to comparison conditions, whereas SCEDs produced a pooled BC-IRR of 4.52 (95% CI [1.71, 11.92]), indicating more than a fourfold increase over baseline. We discuss BC-IRR’s advantages over BC-SMD, namely, its invariance to outcome variance, insensitivity to observation session length, and direct percent-change interpretation, alongside practical limitations, including the requirement of nonzero baselines. These findings illustrate how BC-IRR can support design-comparable, interpretable synthesis of SCED and group-design evidence while underscoring the importance of triangulating ratio-metric and SMD-type effect sizes when evaluating intervention impact.
- 13 July 2026
- Published by American Psychological Association (APA) in Psychological Methods
Bayesian marginalized zero-inflated Poisson model with random effects for single-case experimental designs: A simulation study.
Analyzing zero-inflated count data in single-case experimental designs presents a significant analytical challenge. Although traditional zero-inflated generalized linear mixed models are available, they estimate a conditional treatment effect given an individual coming from the count process, which often mismatches the applied researcher's interest in the overall effect of an intervention for each subject. These conditional models also face interpretational and estimation challenges within the small-sample context of single-case experimental designs. This study introduces and evaluates a Bayesian marginalized zero-inflated Poisson (mZIP) model with random effects. This framework reparameterizes the model to estimate the marginal intervention effect, aligning the statistical estimand with the typical research question. A Monte Carlo simulation study was conducted to evaluate the mZIP model's performance. We compared its performance with other methods that also target the marginal treatment effect: the log response ratio (LRR), a Poisson generalized linear mixed-effects model (GLMM), and a negative binomial GLMM. Simulation results indicate that the Bayesian mZIP model consistently recovers unbiased estimates of the marginal treatment effect and provides reliable statistical inference. The LRR, Poisson GLMM, and negative binomial GLMM produced biased estimates for the marginal effect under conditions with smallest sample size and highest zero-inflation rate. They also suffered from low coverage rate, making their inferential statistics invalid. The LRR indicated lower statistical power than the mZIP model. We apply the mZIP model to two empirical data sets to illustrate its application and interpretation. Finally, we discuss the distinction between conditional and marginal estimands, as well as the limitations and future directions. (PsycInfo Database Record (c) 2026 APA, all rights reserved).
- 13 July 2026
- Published by American Psychological Association (APA) in Psychological Methods
Bayesian marginalized zero-inflated Poisson model with random effects for single-case experimental designs: A simulation study.
Analyzing zero-inflated count data in single-case experimental designs presents a significant analytical challenge. Although traditional zero-inflated generalized linear mixed models are available, they estimate a conditional treatment effect given an individual coming from the count process, which often mismatches the applied researcher's interest in the overall effect of an intervention for each subject. These conditional models also face interpretational and estimation challenges within the small-sample context of single-case experimental designs. This study introduces and evaluates a Bayesian marginalized zero-inflated Poisson (mZIP) model with random effects. This framework reparameterizes the model to estimate the marginal intervention effect, aligning the statistical estimand with the typical research question. A Monte Carlo simulation study was conducted to evaluate the mZIP model's performance. We compared its performance with other methods that also target the marginal treatment effect: the log response ratio (LRR), a Poisson generalized linear mixed-effects model (GLMM), and a negative binomial GLMM. Simulation results indicate that the Bayesian mZIP model consistently recovers unbiased estimates of the marginal treatment effect and provides reliable statistical inference. The LRR, Poisson GLMM, and negative binomial GLMM produced biased estimates for the marginal effect under conditions with smallest sample size and highest zero-inflation rate. They also suffered from low coverage rate, making their inferential statistics invalid. The LRR indicated lower statistical power than the mZIP model. We apply the mZIP model to two empirical data sets to illustrate its application and interpretation. Finally, we discuss the distinction between conditional and marginal estimands, as well as the limitations and future directions. (PsycInfo Database Record (c) 2026 APA, all rights reserved).
- Open Access
- 31 March 2026
- Published by Springer Nature in Research Square
Defining and Estimating Causal Effects in Randomized Alternating Treatment Design for Single-Case Experiments: A Counterfactual Approach
Alternating Treatment Design (ATD) is a commonly used single-case experimental design (SCED) to evaluate the effectiveness of interventions or compare multiple treatments. Commonly used treatment effect estimators, such as simple mean differences and non-overlap indices, are not formally grounded in counterfactual logic and therefore lack a clear causal interpretation. This study addresses this gap by introducing a class of causal estimands in ATDs within a counterfactual framework and the inverse probability weighting (IPW) estimators and augmented IPW (AIPW) estimators for the average immediate treatment effect and the average lag-1 effect. Through Monte Carlo simulation studies, we evaluated the performance of IPW and AIPW estimators under conditions representative of psychological and behavioral research (e.g., short series lengths and restricted randomization). The main simulations compared these methods with commonly used estimators in terms of bias, type I error, and statistical power of randomization tests. The supplementary simulations further assessed the accuracy of standard error estimates and the coverage of corresponding confidence intervals. Results indicate that the AIPW estimator is a promising approach for quantifying the average immediate effect in randomized ATDs, as it produces accurate estimates even in the presence of residual effects and requires fewer assumptions than traditional methods. In addition, the AIPW estimator of the average lag-1 effect provides a practical tool for diagnosing potential multiple treatment interference. Finally, we illustrate the application of IPW and AIPW estimators using real data from an ATD study.
- 31 March 2026
- Published by Springer Nature in Research Square
Open AccessDefining and Estimating Causal Effects in Randomized Alternating Treatment Design for Single-Case Experiments: A Counterfactual Approach
Alternating Treatment Design (ATD) is a commonly used single-case experimental design (SCED) to evaluate the effectiveness of interventions or compare multiple treatments. Commonly used treatment effect estimators, such as simple mean differences and non-overlap indices, are not formally grounded in counterfactual logic and therefore lack a clear causal interpretation. This study addresses this gap by introducing a class of causal estimands in ATDs within a counterfactual framework and the inverse probability weighting (IPW) estimators and augmented IPW (AIPW) estimators for the average immediate treatment effect and the average lag-1 effect. Through Monte Carlo simulation studies, we evaluated the performance of IPW and AIPW estimators under conditions representative of psychological and behavioral research (e.g., short series lengths and restricted randomization). The main simulations compared these methods with commonly used estimators in terms of bias, type I error, and statistical power of randomization tests. The supplementary simulations further assessed the accuracy of standard error estimates and the coverage of corresponding confidence intervals. Results indicate that the AIPW estimator is a promising approach for quantifying the average immediate effect in randomized ATDs, as it produces accurate estimates even in the presence of residual effects and requires fewer assumptions than traditional methods. In addition, the AIPW estimator of the average lag-1 effect provides a practical tool for diagnosing potential multiple treatment interference. Finally, we illustrate the application of IPW and AIPW estimators using real data from an ATD study.
- 23 February 2026
- Published by Springer Nature in Behavior Research Methods
- Vol. 58 (3) , 72
Generalized least squares transformation for single-case experimental design: Introducing the R package lmeSCED
Data generated from single-case experimental designs (SCEDs) are repeated observations on one or a few participants, making multilevel models (MLMs) a usef
- 23 February 2026
- Published by Springer Nature in Behavior Research Methods
- Vol. 58 (3) , 72
Generalized least squares transformation for single-case experimental design: Introducing the R package lmeSCED
Data generated from single-case experimental designs (SCEDs) are repeated observations on one or a few participants, making multilevel models (MLMs) a usef
- Open Access
- 24 October 2025
- Published by Center for Open Science in OSF Preprints
Between-case Incidence Rate Ratio: A Design Comparable Effect Size for Count Outcomes in Single Case Experimental Designs
Single-case experimental designs (SCEDs) are essential for evaluating interventions inpsychological, educational, and behavioral research. While the field has made significantprogress in developing effect size measures for SCEDs, existing metrics, such as the between-case standardized mean difference (BC-SMD), have limitations when applied to count outcomes.Count data often violate assumptions of normality and homoscedasticity, presenting challengesfor accurate effect size estimation. This study addresses the need for an alternative effect sizemeasure tailored to count outcomes in SCEDs that is directly comparable to the incidence rateratio (IRR) used in between-subjects designs. We define the between-case incidence rate ratio(BC-IRR) within a counterfactual framework and demonstrate its equivalence to the IRR inrandomized controlled trials under specific conditions. Using real-world SCED data, weillustrate the estimation of BC-IRR via generalized linear mixed models (GLMMs). The studyalso discusses the assumptions underlying BC-IRR, its potential misuse, and its limitations whileproviding directions for future research.
- 24 October 2025
- Published by Center for Open Science in OSF Preprints
Open AccessBetween-case Incidence Rate Ratio: A Design Comparable Effect Size for Count Outcomes in Single Case Experimental Designs
Single-case experimental designs (SCEDs) are essential for evaluating interventions inpsychological, educational, and behavioral research. While the field has made significantprogress in developing effect size measures for SCEDs, existing metrics, such as the between-case standardized mean difference (BC-SMD), have limitations when applied to count outcomes.Count data often violate assumptions of normality and homoscedasticity, presenting challengesfor accurate effect size estimation. This study addresses the need for an alternative effect sizemeasure tailored to count outcomes in SCEDs that is directly comparable to the incidence rateratio (IRR) used in between-subjects designs. We define the between-case incidence rate ratio(BC-IRR) within a counterfactual framework and demonstrate its equivalence to the IRR inrandomized controlled trials under specific conditions. Using real-world SCED data, weillustrate the estimation of BC-IRR via generalized linear mixed models (GLMMs). The studyalso discusses the assumptions underlying BC-IRR, its potential misuse, and its limitations whileproviding directions for future research.
- Open Access
- 17 October 2025
- Published by Center for Open Science in PsyArXiv
Between-case Incidence Rate Ratio for Single Case Experimental Designs with Count Outcomes: A Monte Carlo Simulation with Conditional and Marginal Models
Publication Update (October 2026): A substantially revised and peer-reviewed version of this manuscript has been published in Research Synthesis Methods. Readers are encouraged to consult and cite the final published version: https://doi.org/10.1017/rsm.2026.10126Calculating design-comparable effect sizes in single-case experimental designs (SCEDs) is essential for research synthesis to identify evidence-based practices that integrate findings from both SCEDs and group-based designs. The field has made significant progress in developing additive design-comparable effect sizes in recent years. This study aims to evaluate the statistical properties of a newly developed proportional estimator, the between-case incidence rate ratio (BC-IRR), for SCEDs with count outcomes. Two analytical frameworks are introduced for estimating BC-IRR: generalized linear mixed models (GLMMs) and generalized estimating equations (GEEs). A large-scale Monte Carlo simulation is conducted to assess the bias of point estimate, the standard error bias, mean squared error, and coverage rate. We also demonstrate the estimation of BC-IRR using GLMMs and GEEs with real data. Based on the simulation and demonstration results, we provide recommendations for applied researchers, discuss limitations, and outline directions for future research.
- 17 October 2025
- Published by Center for Open Science in PsyArXiv
Open AccessBetween-case Incidence Rate Ratio for Single Case Experimental Designs with Count Outcomes: A Monte Carlo Simulation with Conditional and Marginal Models
Publication Update (October 2026): A substantially revised and peer-reviewed version of this manuscript has been published in Research Synthesis Methods. Readers are encouraged to consult and cite the final published version: https://doi.org/10.1017/rsm.2026.10126Calculating design-comparable effect sizes in single-case experimental designs (SCEDs) is essential for research synthesis to identify evidence-based practices that integrate findings from both SCEDs and group-based designs. The field has made significant progress in developing additive design-comparable effect sizes in recent years. This study aims to evaluate the statistical properties of a newly developed proportional estimator, the between-case incidence rate ratio (BC-IRR), for SCEDs with count outcomes. Two analytical frameworks are introduced for estimating BC-IRR: generalized linear mixed models (GLMMs) and generalized estimating equations (GEEs). A large-scale Monte Carlo simulation is conducted to assess the bias of point estimate, the standard error bias, mean squared error, and coverage rate. We also demonstrate the estimation of BC-IRR using GLMMs and GEEs with real data. Based on the simulation and demonstration results, we provide recommendations for applied researchers, discuss limitations, and outline directions for future research.
- 11 January 2025
- Published by Wiley in British Journal of Health Psychology
- Vol. 30 (1) , e12781
Pre‐disability resilience and well‐being following the onset of functional impairments
Objectives: Tested the prospective relationship of a resilient personality prototype determined prior to disability onset to well‐being among persons with and without debilitating functional impairments nine to 10 years later. A resilient profile was expected to predict well‐being through its beneficial associations with positive affect, perceived control and social support. Design: Longitudinal, prospective observation study. Methods: Data obtained from participants with no functional impairments at the first assessment of the Midlife in the United States (MIDUS) project and who participated in the second survey (1147 men, 985 women at Time 1). Cluster analysis was used to create personality profiles at the first time point. Differences between resilient and non‐resilient individuals were examined. A structural equation model (SEM) tested the prospective effects of resilience on positive affect, perceived control and social support to well‐being. Results: A resilient personality profile was identified, as expected ( n = 877). Individuals with a resilient personality prototype reported higher social support, positive affect, perceived control, life satisfaction and self‐rated health at both measurement occasions than the non‐resilient group. The SEM revealed that a resilient prototype operated through the three mediating variables to prospectively predict life satisfaction, and through social support and positive affect to predict self‐rated health. These effects were independent of gender and disability severity. Conclusions: A resilient personality profile prospectively operates through positive affect, perceived control and social support to predict well‐being following the onset of debilitating impairments. The theoretical and clinical implications of these findings are discussed, and limitations are considered.
- 11 January 2025
- Published by Wiley in British Journal of Health Psychology
- Vol. 30 (1) , e12781
Pre‐disability resilience and well‐being following the onset of functional impairments
Objectives: Tested the prospective relationship of a resilient personality prototype determined prior to disability onset to well‐being among persons with and without debilitating functional impairments nine to 10 years later. A resilient profile was expected to predict well‐being through its beneficial associations with positive affect, perceived control and social support. Design: Longitudinal, prospective observation study. Methods: Data obtained from participants with no functional impairments at the first assessment of the Midlife in the United States (MIDUS) project and who participated in the second survey (1147 men, 985 women at Time 1). Cluster analysis was used to create personality profiles at the first time point. Differences between resilient and non‐resilient individuals were examined. A structural equation model (SEM) tested the prospective effects of resilience on positive affect, perceived control and social support to well‐being. Results: A resilient personality profile was identified, as expected ( n = 877). Individuals with a resilient personality prototype reported higher social support, positive affect, perceived control, life satisfaction and self‐rated health at both measurement occasions than the non‐resilient group. The SEM revealed that a resilient prototype operated through the three mediating variables to prospectively predict life satisfaction, and through social support and positive affect to predict self‐rated health. These effects were independent of gender and disability severity. Conclusions: A resilient personality profile prospectively operates through positive affect, perceived control and social support to predict well‐being following the onset of debilitating impairments. The theoretical and clinical implications of these findings are discussed, and limitations are considered.
- Open Access
- 22 December 2023
- Published by Leibniz Institute for Psychology (ZPID) in Methodology
- Vol. 19 (4) , 365-386
A comparison of methods for specifying optimal random effects structures
Using Monte Carlo simulations, this study compared the performance of various approaches to the specification of random effects structures in linear mixed effects models (LMMs), including the minimal approach, the maximal approach, the forward search, the backward search, and the all-possible structures approach. The results showed that if the predictor of interest is at the within-cluster level or involves a cross-level interaction, the maximal approach, the best-path forward search, and the best-path backward search are all desirable methods. If the predictor of interest is at the cluster level, it is not essential to specify random slopes of Level-1 predictors. In addition, it is important to specify random slopes of within-cluster control variables, as they can increase the statistical power for testing the main within-cluster variables, especially when the sample size is small and the variance of the random slope of the control variable is large.
- 22 December 2023
- Published by Leibniz Institute for Psychology (ZPID) in Methodology
- Vol. 19 (4) , 365-386
Open AccessA comparison of methods for specifying optimal random effects structures
Using Monte Carlo simulations, this study compared the performance of various approaches to the specification of random effects structures in linear mixed effects models (LMMs), including the minimal approach, the maximal approach, the forward search, the backward search, and the all-possible structures approach. The results showed that if the predictor of interest is at the within-cluster level or involves a cross-level interaction, the maximal approach, the best-path forward search, and the best-path backward search are all desirable methods. If the predictor of interest is at the cluster level, it is not essential to specify random slopes of Level-1 predictors. In addition, it is important to specify random slopes of within-cluster control variables, as they can increase the statistical power for testing the main within-cluster variables, especially when the sample size is small and the variance of the random slope of the control variable is large.