{"id":{"repo_id":"msu","oai_identifier":"oai:d.lib.msu.edu:etd_52625"},"canonical_url":"https://search.ndltd.org/etd/msu/oai:d.lib.msu.edu:etd_52625","repository":{"repo_id":"msu","name":"Michigan State University","base_url":"https://d.lib.msu.edu/oai"},"display":{"title":"Two-Level Hurdle Model of the GPA Distribution with Two-Side Truncated Generalized Poisson Regression","abstract":"Thesis (Ph.D.)--Michigan State University. Measurement and Quantitative Methods - Doctor of Philosophy, 2025","abstract_html":"Thesis (Ph.D.)--Michigan State University. Measurement and Quantitative Methods - Doctor of Philosophy, 2025","abstract_has_math":false,"creators":["Liu, Jiachen"],"institution":null,"degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":["Schmidt, William H","Schneider, Barbara L","Houang, Richard T","Cui, Yuehua","Weng, Haolei"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2025,"date_issued":"2025","date_published":"2025","updated_at":"2026-11-18T18:59:09Z","subjects":["Educational evaluation","Statistics","Educational tests & measurements","Educational tests and measurements"],"languages":["eng"],"rights":["Attribution-NonCommercial-NoDerivatives 4.0 International"],"rights_urls":[],"identifier_entries":[{"key":"dc:identifier","label":"Identifier","values":["etd:52625","local:Liu_grad.msu_0128D_21591"],"render_values":[{"text":"etd:52625","href":null,"code":true},{"text":"local:Liu_grad.msu_0128D_21591","href":null,"code":true}]}],"person_identifier_entries":[]},"links":{"outbound_url":"https://doi.org/doi:10.25335/3141-k074","outbound_label":"DOI","outbound_source":"dc:identifier","record_url":"https://doi.org/doi:10.25335/3141-k074","record_label":"Repository record","record_kind":"doi","record_source":"dc:identifier","download_url":null,"download_label":null,"download_source":null,"access_status":null,"embargo_until":null,"access_note":null},"source_record":{"url":"https://d.lib.msu.edu/oai?verb=GetRecord&metadataPrefix=oai_dc&identifier=oai%3Ad.lib.msu.edu%3Aetd_52625","prefix":"oai_dc"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Schmidt, William H","Schneider, Barbara L","Houang, Richard T","Cui, Yuehua","Weng, Haolei"]},{"key":"dc:creator","label":"Author","values":["Liu, Jiachen"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2025"]},{"key":"dc:relation","label":"Dc Relation","values":["Electronic Theses & Dissertations"]},{"key":"dc:type","label":"Dc Type","values":["text","Electronic dissertations"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Educational evaluation","Statistics","Educational tests & measurements","Educational tests and measurements"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["eng"]},{"key":"dc:rights","label":"Dc Rights","values":["Attribution-NonCommercial-NoDerivatives 4.0 International"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["etd:52625","local:Liu_grad.msu_0128D_21591","https://doi.org/doi:10.25335/3141-k074"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Thesis (Ph.D.)--Michigan State University. Measurement and Quantitative Methods - Doctor of Philosophy, 2025","Grade Point Average (GPA) has been a key indicator of academic performance within educational research and policy. The historical development of GPA scores in the United States went through four important stages: the imitation stage for the introduction of categorical grading system from Europe, the systematization stage that build up the foundation for communicating merit both within and outside of schools, the standardization stage where reliability and validity concerns were raised, and the resistance stage when grade inflation and rescaling of scores become prevalent (Brookhart et al., 2016; Schneider & Hutt, 2014). These transitions charted the trajectory of how student GPA evolved from a simple communicative tool to a contested measure of merit and ability, calling for a closer examination.Despite the widespread use of GPA scores, a lack of universal grading rubric from teachers and therefore a consistent pattern in the GPA distributions posed real challenges in modeling these scores. This dissertation builds on the framework of the three properties of the normal distribution- unboundedness, symmetry, and unimodality- and explores alternative distributional assumptions and modeling techniques, such as beta and logit-normal distributions, log transformations, and zero-inflated or multimodal models, to address violations of these properties. This dissertation took the opportunity to re-analyze the National Study of Learning Mindsets (NSLM) dataset and modified the two-level hurdle model to account for the upper bound of GPA scores, as well as zero-inflation and under-dispersion in the conditional distribution. These modifications would be necessary in that failure to account for zero-inflation and under-dispersion often lead to biased estimates and inflated standard errors and therefore misleading conclusions. These models were also applied to study the effect of textbooks as contextual environment in sustaining and perpetuating growth mindset effects.Although Conway-Maxwell-Poisson (CMP) and generalized Poisson (GP) can both be applied to account for under-dispersion in the distribution, the CMP distribution does not directly link its mean with the covariates, unlike the GP distribution. The minimization-maximization algorithm for stable estimation of the parameters in GP regression was proposed by Li et al. (2023). However, this algorithm does not deal with cases when random effects are involved. This dissertation turned to the R packages glmmTMB and RTMB for parameter estimation. These two packages effectively estimate parameters in GLMMs using Laplace approximation for integration, Newton’s method for inner optimization, and Broyden-Fletcher-Godfarb-Shanno (BFGS) Method for outer optimization. Findings in this dissertation revealed a different story for the growth mindset intervention. Through the textbook analysis, specific types of instructional practices were recommended in helping to improve student grades in mathematics and reinforcing the growth mindset effect. Differences in the conclusions drawn from different distributional assumptions implies the importance of aligning modeling techniques with the real-world data. While GPA has drawn criticism for its limitations in evaluating merit and progress, particularly in the context of grade inflation, it remains a widely used and straightforward indicator of academic performance. When interpreted through appropriate methodological frameworks, it can meaningfully reflect student growth. As such, it is essential that researchers model GPA outcomes with rigor, nuance, and sensitivity to context to ensure fair and valid insights into student achievement.","Description based on online resource. Title from PDF t.p. (Michigan State University Fedora Repository, viewed ).","Includes bibliographical references."]},{"key":"dc:format","label":"Dc Format","values":["application/pdf","1 online resource (186 p.) : ill."]},{"key":"dc:title","label":"Title","values":["Two-Level Hurdle Model of the GPA Distribution with Two-Side Truncated Generalized Poisson Regression"]}]}],"canonical_facts":{"dc:contributor":["Schmidt, William H","Schneider, Barbara L","Houang, Richard T","Cui, Yuehua","Weng, Haolei"],"dc:creator":["Liu, Jiachen"],"dc:date":["2025"],"dc:description":["Thesis (Ph.D.)--Michigan State University. Measurement and Quantitative Methods - Doctor of Philosophy, 2025","Grade Point Average (GPA) has been a key indicator of academic performance within educational research and policy. The historical development of GPA scores in the United States went through four important stages: the imitation stage for the introduction of categorical grading system from Europe, the systematization stage that build up the foundation for communicating merit both within and outside of schools, the standardization stage where reliability and validity concerns were raised, and the resistance stage when grade inflation and rescaling of scores become prevalent (Brookhart et al., 2016; Schneider & Hutt, 2014). These transitions charted the trajectory of how student GPA evolved from a simple communicative tool to a contested measure of merit and ability, calling for a closer examination.Despite the widespread use of GPA scores, a lack of universal grading rubric from teachers and therefore a consistent pattern in the GPA distributions posed real challenges in modeling these scores. This dissertation builds on the framework of the three properties of the normal distribution- unboundedness, symmetry, and unimodality- and explores alternative distributional assumptions and modeling techniques, such as beta and logit-normal distributions, log transformations, and zero-inflated or multimodal models, to address violations of these properties. This dissertation took the opportunity to re-analyze the National Study of Learning Mindsets (NSLM) dataset and modified the two-level hurdle model to account for the upper bound of GPA scores, as well as zero-inflation and under-dispersion in the conditional distribution. These modifications would be necessary in that failure to account for zero-inflation and under-dispersion often lead to biased estimates and inflated standard errors and therefore misleading conclusions. These models were also applied to study the effect of textbooks as contextual environment in sustaining and perpetuating growth mindset effects.Although Conway-Maxwell-Poisson (CMP) and generalized Poisson (GP) can both be applied to account for under-dispersion in the distribution, the CMP distribution does not directly link its mean with the covariates, unlike the GP distribution. The minimization-maximization algorithm for stable estimation of the parameters in GP regression was proposed by Li et al. (2023). However, this algorithm does not deal with cases when random effects are involved. This dissertation turned to the R packages glmmTMB and RTMB for parameter estimation. These two packages effectively estimate parameters in GLMMs using Laplace approximation for integration, Newton’s method for inner optimization, and Broyden-Fletcher-Godfarb-Shanno (BFGS) Method for outer optimization. Findings in this dissertation revealed a different story for the growth mindset intervention. Through the textbook analysis, specific types of instructional practices were recommended in helping to improve student grades in mathematics and reinforcing the growth mindset effect. Differences in the conclusions drawn from different distributional assumptions implies the importance of aligning modeling techniques with the real-world data. While GPA has drawn criticism for its limitations in evaluating merit and progress, particularly in the context of grade inflation, it remains a widely used and straightforward indicator of academic performance. When interpreted through appropriate methodological frameworks, it can meaningfully reflect student growth. As such, it is essential that researchers model GPA outcomes with rigor, nuance, and sensitivity to context to ensure fair and valid insights into student achievement.","Description based on online resource. Title from PDF t.p. (Michigan State University Fedora Repository, viewed ).","Includes bibliographical references."],"dc:format":["application/pdf","1 online resource (186 p.) : ill."],"dc:identifier":["etd:52625","local:Liu_grad.msu_0128D_21591","https://doi.org/doi:10.25335/3141-k074"],"dc:language":["eng"],"dc:relation":["Electronic Theses & Dissertations"],"dc:rights":["Attribution-NonCommercial-NoDerivatives 4.0 International"],"dc:subject":["Educational evaluation","Statistics","Educational tests & measurements","Educational tests and measurements"],"dc:title":["Two-Level Hurdle Model of the GPA Distribution with Two-Side Truncated Generalized Poisson Regression"],"dc:type":["text","Electronic dissertations"]},"updated_at":"2026-11-18T18:59:09Z"}