Amota Ataneka

 Amota Ataneka Merang is a doctoral candidate in Quantitative Research Methodologies specializing in causal machine learning and psychometrics. His dissertation develops new algorithms for causal inference with latent variables — constructs that are important to us because they influence policy, practice, and theory development but cannot be measured directly (e.g., depression, burnout, resilience, emotions, affect, satisfaction, student achievement, motivation, pain severity, social capital, self-efficacy, mobility, trust, belonging, stigma, food insecurity, quality of life, organizational culture, radicalization, etc.). Such constructs are ubiquitous across psychology, education, health sciences, economics, political science, public health, climate science, and beyond, appearing as both outcomes and predictors in the same study. For example, a researcher estimating the causal effect of a treatment on depression must often account for prior mental health conditions (e.g., anxiety and depressive symptoms) as confounders — all of which are latent (i.e., they cannot be directly measured the way height or weight can). 

Despite their pervasiveness across research disciplines, existing causal inference methods handle latent variables poorly: traditional and contemporary approaches (e.g., Causal Forests, Structural Equation Modeling) either ignore measurement error entirely or impose strong parametric assumptions about functional form, requiring that all nonlinear terms and interaction effects be correctly specified in advance. In practice, these assumptions are routinely violated or remain unverifiable because the true relationships among observed and latent variables are complex and largely unknown. Machine learning and data-adaptive methods relax these model specification constraints, but they do so while ignoring the measurement error inherent in latent constructs. Bridging statistics, machine learning, and quantitative methodology, Amota's work addresses both of these limitations simultaneously. The new machine learning methods he is developing account for measurement error and relax model specification assumptions, enabling cause-and-effect analysis with latent variables in nonexperimental and observational settings, even when the functional relationships among variables are unknown or complex. The result methods integrate psychometric theory, causal inference, and machine learning — one that simultaneously accommodates the flexibility needed to capture complex relationships and the rigor needed to handle constructs that are measured with error. This dissertation work is being supported ($27,500) by the National Academy of Education/Spencer Dissertation Fellowship, the most prestigious dissertation fellowship in the field of Education. 

A second line of his research work addresses the development and design of adequately powered studies when researchers care about not only main effects but also indirect effects in multilevel and multisite settings. This research directly responds to a well-documented problem in educational, health, psychotherapy research: investigators routinely underpowered their main and mediation effect studies because they apply single-level power formulas to clustered data, or rely on standardized conventions (e.g., Cohen's d) rather than empirically grounded design parameters. His peer-reviewed publications in this area include: (1) Designing Multisite Randomized Trials to Detect (Conditional) Indirect Effects (American Journal of Evaluation, 2026, (2) Design and Analysis of Multisite Cluster-Randomized Trials Targeting (Conditional) Mediation Effects (Journal of Experimental Education, 2025), (3) Design Parameter Values for Planning Mediation Studies with Teacher and Student Mathematics Outcomes (Journal of Research on Educational Effectiveness, 2024), and (4) Evaluations of Literacy-Based Programs: Empirical Values for Designing Studies Probing Mediation (Evaluation Review, 2026). These papers developed methods (principles and expressions) to predict statistical power and sample size in complex settings specific for mediation effects and provided empirically derived design parameter values (drawn from real large-scale datasets) for such settings. A related work is providing PowerUpR Shiny App to implement power analysis for standard and complex data structures (see Causal Evaluation & Ataneka et al., 2023).

A third line of his research agenda focuses on QuantCrit and the development of equitable quantitative methodologies for evaluating institutions and interventions serving remote and historically marginalized communities. This work is deeply connected to his personal journey from subsistence living in Nikunau, a very remote atoll in Kiribati, to doctoral training in the United States through formal education. Growing up in a context where conventional indicators of “success” often failed to capture local realities shaped his interest in questioning how quantitative systems define disadvantage, achievement, and institutional performance. In this area, he developed Critical Data Envelopment Analysis (Critical DEA), a QuantCrit framework for evaluating homogeneous entities such as schools, hospitals, banks, and ports in ways that center equity, local context, and community-defined strengths.

Amota is also writing a book chapter for the Oxford Handbook of Impact Evaluation and regularly presents his work at premier research conferences: American Educational Research Association (AERA), Modern Modeling Methods (M3), the American Evaluation Association (AEA), and Society for Research on Educational Effectiveness (SREE). 

Bachelor of Economics: University of Queensland Australia, 2011

Master's of Public Administration (MPA): University of Hawaii USA, 2017 (Data Envelopment Analysis)

Ph.D. (Quantitative Methodologies): University of Cincinnati Ohio, 2027 (Causal Machine Learning, Psychometrics)

Ph.D. (Education): James Cook University Australia, 2025 (Public Education Administration)

- Teaching Assistant: Graduate-level Research Methods Courses: Course Title: EDST8086 Causal Inference for the Social Sciences; Where: Inter-university Consortium for Political and Social Research (ICPSR), Institute for Social Research (ISR), University of Michigan, USA; When: Summer 2024 ,

- Teaching Assistant: Graduate-level Research Methods Courses: Course Title: EDST8086 Experimental & Quasi-experimental Research for Causal Inference Where: School of Education, College of Education, Criminal Justice and Human Services, University of Cincinnati, Ohio, USA. When: Fall 2023; Spring 2024; Fall 2024,

- Teaching Assistant: Graduate-level Research Methods Courses: Course Title: EDST8089 Structural Equation Modeling Where: School of Education, College of Education, Criminal Justice and Human Services, University of Cincinnati, Ohio, USA When: Spring 2025,

- Workshop/Professional-Development Instructor:1. Workshop Name: PD25-10 Designing Adequately Powered Cluster and Multisite Randomized Trials to Detect Main Effects, Moderation, and Mediation Where: American Educational Research Association (AERA) Conference, Denver, Colorado, USA. Link When: April 2025, https://tinyurl.com/28cpmdee

- Workshop/Professional-Development Instructor:Workshop Name: Designing Adequately Powered Cluster and Multisite Randomized Trials to Detect Main Effects, Moderation, and Mediation. Where: American Educational Research Association (AERA) Conference, Chicago, IL, USA. Link When: April 2023 Notes: Participants paid $90 on top of conference registration fee to attend this workshop,

- Workshop/Professional-Development Instructor:Workshop Name: Designing Adequately Powered Cluster and Multisite Randomized Trials to Detect Main Effects, Moderation, and Mediation. Where: Annual Meeting of the American Evaluation Association (AEA), Indianapolis, USA. When: October 2023 Notes: Paid workshop; Link , https://eval23.eventscribe.net/fsPopup.asp?PresentationID=1282613&mode=presInfo

- Workshop/Professional-Development Instructor:Workshop Name: Designing Adequately Powered Multilevel Studies Cluster Randomized Trials to Detect Main Effects, Moderation, and Mediation. Where: Northeastern Educational Research Conference (NERA) Conference, Trumbull, Connecticut, USA. Link When: October 2023 Notes: Paid workshop, https://www.nera-education.org/annual_conference.php,

Grant: #1760884 Investigators:PIs: Benjamin Kelcey, Nianbo Dong National Science Foundation (NSF) Designing Multisite Mediation Studies to Track Teacher Development Processes in Mathematics Role:Collaborator $499,996.00 Completed Type:Grant

Grant: #1552535 Investigators:PIs: Benjamin Kelcey National Science Foundation (NSF) Multilevel Mediation Models to Study the Impact of Teacher Development on Student Achievement in Mathematics. Role:Contributor $679,105.00

Grant: #2321191 01-01-2024 -12-31-2026 U.S National Science Foundation (NSF) An Explanatory Machine Learning Framework for Teacher Effectiveness in STEM Education Role:Contributor $349,999.00 Active Type:Grant

Investigators:McLean LE, Youngs P, Bartell T, Kelcey B, Jones ND The Impacts of Pre-service Supervised Field Experiences on Elementary Teachers Retention and Effectiveness in Mathematics Role:Contributor $1,493,082.00 Type:Grant

Peer Reviewed Publications

Bai, F., Xie, Y., Kelcey, B., Ataneka, A., McLean, L., & Phelps, G. (2024). Design Parameter Values for Planning Mediation Studies with Teacher and Student Mathematics Outcomes. Journal of Research on Educational Effectiveness, 1–35. https://doi.org/10.1080/19345747.2024.2349670

Ataneka, A., Bai, F., Xie, Y., Kelcey, B., McLean, L. (2026. ) Evaluations of Literacy-Based Programs: Empirical Values for Designing Studies Probing Mediation.Eavaluation Review, , 0 (0 ) , More Information

Ataneka, A., Doyle, T., Tomas, L., (in progress). ‘Red-Dirt Thinking’ of School Performance Evaluation: A CritQuant Application of Data Envelopment Analysis (DEA) in Queensland State Primary Schools.

Bai, F., Kelcey, B., Ataneka, A., Xie, Y., Cox, K., & Dong, N. (2025). Designing Multisite Randomized Trials to Detect (Conditional) Mediation Effects. American Journal of Evaluation. More Information

Bai, F., Kelcey, B., Xie, Y., Ataneka, A., Cox, K., & Dong, N. (2025. ) Design and Analysis of Multisite Cluster-Randomized Trials Targeting (Conditional) Mediation Effects.The Journal of Experimental Education , , More Information

Bai, F., Kelcey, B., Ataneka, A., (in-progress), Missing Data in Multilevel Structural Equation Models: A Multilevel Structural-After-Measurement Estimation Approach, Link

Ataneka, A., Kelcey, B., McLean, L., Youngs, P., (under review), Evaluating the Causal Effect of Teacher Mental Health Induction Program on Teacher Burnout, Depression and Anxiety, Educational Evaluation and Policy Analysis Link

Ataneka, A., Kelcey, B Causal Machine Learning with Latent Outcomes .Psychological Methods, ,

Other Publications

Ataneka, A., Kelcey, B., Dong, N., Bulus, M., & Bai, F. (2023. ) Power-Up Shiny App User Guide 0.9 .https://www.causalevaluation.org/power-analysis.html,

Ataneka A., (2016. ) Evaluating the efficiency of public hospitals in the State of Hawaii .

Hawaii public Hospitals Efficiency

Book Chapter

Ataneka A., Kelcey B., (2026 ) Cluster Randomized Trials (CRT) Oxford Handbook for Impact Evaluations .Oxford Press

Invited Presentations

Bai, F., Ataneka, A., Kelcey, B., & Xie, Y (2023. ) Designing Adequately Powered Multilevel Studies Cluster Randomized Trials to Detect Main Effects, Moderation, and Mediation .University of Connecticut, Trumbull, Connecticut, USA, https://www.nera-education.org/annual_conference.php. Workshop. . Level:Regional

Ataneka, A., Xie, Y., Bai, F., Kelcey, B., & Dong, N. (Oct 2023). Symposium Title: Exploring Variation in Populations and Outcomes: Challenging Design Conventions in Study Parameters, Paper Presented: Design Parameter Values for Multilevel Mediation Studies of Teacher Development. Northeastern Educational Research Conference (NERA) Conference 2023, Trumbull, Connecticut, USA. Link

Xie, Y., Bai, F., Ataneka, A., & Kelcey, B., (2024), Symposium Title: Design and Analysis for Causal Inferences in and across, Paper Presented: Moderation in Multisite Partially Nested Trials: Estimation, Inference, and Design Studies, Baltimore, Maryland, USA, Link

Poster Presentations

Bai, F., Kelcey, B., Ataneka, A., Xie, Y., Dong, N., (April 2025), Experimental Power for Mediation Ataneka, A., Xie, Y., Bai, F., Kelcey, B., & Dong, N. (Oct 2023). Design Parameter Values for Multilevel Mediation Studies of Teacher Development. Effects in Multisite Studies, Northeastern Educational Research Association (NERA), Tru Colorado, USA. Link

Colloquium

Ataneka, A., Xie, Y., Bai, F., Kelcey, B., & Dong, N. (04-2024). Design Parameter Values for Planning Mediation Studies With Teacher and Student Mathematics Outcomes. . Spring Conference, Louisville, Kentucky. Conference. Level:Regional

Paper Presentations

Ataneka, A., Xie, Y., Bai, F., Kelcey, B., & Dong, N. (05-2023. ) Design Parameter Values for Planning Mediation Studies With Teacher and Student Mathematics Outcomes. .Chicago. Conference. Level:National

Ataneka, A., (2017. ) Examining the Efficiency of Public Elementary Schools in the State of Hawaii: An Application of Data Envelopment Analysis .California. Conference. Level:National

Bai, F., Xie, Y., Kelcey, B., Ataneka, A., McLean, L., Dong, N., & Phelps, G. (10-2023. ) Design Parameter Values for Multilevel Mediation Studies of Teacher Development. .Connecticut. Conference.

Ataneka, A., Xie, Y., Bai, F., Kelcey, B., & Dong, N. (09-2023. ) Design Parameter Values for Multilevel Mediation Studies of Teacher Development. .Arlington Virginia . Conference. Level:National

Ataneka, A., Kelcey, B., McLean, L., Youngs, P., (April 2025), Evaluating the Causal Effect of Teacher Mental Health Induction Program on Teacher Burnout, Depression and Anxiety, Presented at the American Educational Research Association (AERA) Conference, Denver, Colorado, USA, Link

Kelcey, B., Ataneka, A., (April 2025), Explanatory Machine Learning Methods for Teacher Effectiveness, American Educational Research Association (AERA) Conference, Denver, Colorado, USA, Link

Ataneka, A., Xie, Y., Bai, F., Kelcey, B., & Dong, N. (Oct 2023). Design Parameter Values for Multilevel Mediation Studies of Teacher Development. Presented at the Annual Meeting of the American Evaluation Association (AEA), Indianapolis, IN. USA, Link

Kelcey, B., Xie, Y., Bai, F., & Ataneka, A. (November 2023). Moderation in Multisite Partially Nested Experiments: Estimation, Inference and Design. Association for Public Policy Analysis & Management (APPAM) Fall Conference 2023, Atlanta, GA, USA.

Kelcey, B., Bai, F., Cox, K., Xie, Y., & Ataneka, A. (April 2024). Structural After Measurement Estimation With Missing Data in Structural Equation Models. American Educational Research Association (AERA) Annual Meeting 2024, Philadelphia, PA, USA. Link

Kelcey, B., Bai, F., & Ataneka, A. (April 2024). Structural After Measurement Estimation of higher order and (non)linear in n-level Structural Equation Models, American Evaluation Association (AEA) Annual Meeting 2024, Indianapolis, Indiana, USA, Link

Kelcey, B., Bai, F., Ataneka, A., & Xie, Y. (April 2024). Experimental Design and Analysis of Multiple Group Individually-Randomized Group Trials. American Educational Research Association (AERA) Annual Meeting 2024, Philadelphia, PA, USA.

Ataneka, A., Doyle, T., Engel, L., (2021), ‘Red-Dirt Thinking’ of School Performance Evaluation: A CritQuant Application of Data Envelopment Analysis (DEA) in Queensland State Primary Schools. CASE HDR Conference, Cairns, Australia

(2026. )

(2026. )

Symposium

Ataneka, A., Cox, K., Kelcey, B., Bai, F., Luce, Hannah., (10-2022. ) Detecting Moderation Effects in Cluster Randomized Trials with Partially Nested Data .Washington DC. Conference. Level:International

Ataneka, A., (Discussant) (11-2022. ) Principles of Estimation, Inference & Design in Partially Nested Regression Discontinuity Studies .New Orleans, Louisiana. Conference.

2022 -2025: Official Member AERA (American Educational Research Association) , Washington DC

2022 -2025: Official Member AEA (American Evaluation Association), Washington DC

2022 -2025: Official Member APPAM (Association for Public Policy Analysis and Management), Washington DC

2022 -2025: Official Member SREE (Society for Research on Educational Effectiveness), Washington DC