Quantitative Psychology: The 86th Annual Meeting of the Psychometric Society, Virtual, 2021 (Springer Proceedings in Mathematics & Statistics, 393)

The volume represents presentations given at the 86th annual meeting of the Psychometric Society, held virtually on July 19–23, 2021. About 500 individuals contributed paper presentations, symposiums, poster presentations, pre-conference workshops, keynote presentations, and invited presentations. Since the 77th meeting, Springer has published the conference proceedings volume from this annual meeting to allow presenters to share their work and ideas with the wider research community, while still undergoing a thorough review process. This proceedings covers a diverse set of psychometric topics, including item response theory, Bayesian models, reliability, longitudinal measures, and cognitive diagnostic models. Preface Contents Some Adventures in Reliability Estimation 1 Background 2 Interdependencies: The Story of Easy or Hard 3 Counterbalancing: The Story of Magnitudes 4 Composite Scores: The Story of Balloons 5 Composite Scores: The Story of Lengths 6 Simulation and Modeling: The Story of Digits 7 Conclusion References Detecting Testlet Effects in Cognitive Diagnosis Models 1 Introduction 2 Cognitive Diagnosis Models 3 Fit Statistics for Local Dependence 4 Simulation Studies 5 Results 5.1 Type I Error Study 5.2 Power Study: Testlet Model 6 Discussion References Comparison of Estimation Algorithms for Latent Dirichlet Allocation 1 Introduction 2 Latent Dirichlet Allocation (LDA) 2.1 Estimation Method 3 Methods 3.1 Simulation Design 3.2 Parameter Estimation 3.3 Label Switching 4 Results 5 Discussion References Relationship Between Students' Test Results and Their Performance in Higher Education Using Different Test Scores 1 Introduction 2 Method 2.1 Test Scores 2.2 The SweSAT 2.3 Participants 2.4 Statistical Analysis 3 Results 4 Discussion References Statistical Properties of Lower Bounds and Factor Analysis Methods for Reliability Estimation 1 Introduction 1.1 CTT Coefficients 1.2 FA Coefficients 2 Simulation Study 2.1 Method Data Generation from GRM Data Generation from FA-Model Conditions Outcome Variables 2.2 Results Unidimensional Models Multidimensional Models Misspecified Models 3 Discussion References Modeling Covarying Responses in Complex Tasks 1 Introduction 2 Methods 2.1 Rasch Model 2.2 Rasch Model with Response-Level Covariates 2.3 Joint Rasch Model 2.4 Model Evaluation 3 Application: Forensic Fingerprint Comparisons 3.1 Model Formulation 4 Results 4.1 Model Comparison 4.2 Parameter Estimates 5 Discussion References Exploring the Utility of Nonfunctional Distractors 1 Methods 2 Results 3 Discussion References Continuation Ratio Model for Polytomous Items Under Complex Sampling Design 1 Introduction 2 Polytomous Scales and Item Response Theory Models 2.1 An Applied Example of Polytomous Scale 2.2 Item Response Models 2.3 Continuation Ratio Model 3 Inferences to the Population 4 Methods 4.1 Selected Data and Measures for Illustrations 4.2 Data Expansion Technique 4.3 Analytical Strategy 5 Results 5.1 Item Side 5.2 Person Side 6 Conclusion and Discussion References Impact of Likelihoods on Class Enumeration in Bayesian Growth Mixture Modeling 1 Introduction 2 Bayesian GMM Model Selection 2.1 A Brief Review of Growth Mixture Models 2.2 Conditional and Marginal Likelihoods of Growth Mixture Models 2.3 Model Comparison Criteria 3 A Simulation Study 3.1 Results 4 Discussion References A Bi-level Individualized Adaptive Learning Recommendation System Based on Topic Modeling 1 Introduction 2 Method 2.1 Bi-term Topic Model 2.2 Loss Function and Gradient Descent 3 Data and Analytic Framework 3.1 Data Description 3.2 Bi-level Recommendation Framework 3.3 Analytic Procedures 4 Results 4.1 Step 1: Feature Learning Component with BTM and JSD 4.2 Step 2: Measurement Component with Gradient Descent 4.3 Step 3: Recommendation Strategy Component with Predicted Total Scores 5 Discussion and Conclusion Appendix Data R Code References Derivation of the Percentile Based Parameters for Tukey HH, HR and HQ Distributions 1 Introduction 2 Tukey g-h Family of Distributions and the Extensions 3 Derivation of the Fifth-Ordered Percentile Based System Parameters 4 Discussion References Predicting Human Psychometric Properties Using Computational Language Models 1 Introduction 2 Background in Natural Language Processing 2.1 Language Modeling 2.2 Natural Language Inference 2.3 Benchmarks of Commonsense Reasoning 3 Gathering Language Model Data 4 Human Studies 5 Experimental Results 5.1 Classical Test Theory 5.2 Item Response Theory 6 Related Work 7 Conclusion References Predicting Item Characteristic Curve (ICC) Using a SoftmaxClassifier 1 Introduction 2 Method 3 Data 4 Results 5 Discussion References Pooled Autoregressive Models for Categorical Data 1 Introduction 2 Autoregressive Model and Categorical Data 2.1 First-Order Autoregressive Model, AR(1) 2.2 Categorical Data 3 Estimation Methods and Likelihoods 3.1 Exact MLE for Pooled Likelihood Function 3.2 Exact MLE for Pooled Data 3.3 Conditional MLE for Pooled Likelihood Function 3.4 Conditional MLE for Pooled Data 4 Simulation Study 4.1 Data Generation 4.2 Model Estimation and Evaluation 5 Results, Conclusions and Discussion 5.1 Results 5.2 Conclusions 5.3 Discussion Appendix 1 Appendix 2 References An Investigation of Prior Specification on Parameter Recovery for Latent Dirichlet Allocation of Constructed-Response Items 1 Introduction 1.1 Purpose of Study 2 Latent Dirichlet Allocation 2.1 Mathematical Formulation 2.2 Generative Model 2.3 Requirements for Analysis 3 Simulation Design 3.1 Simulation Design and Conditions 3.2 Data Generation 3.3 Model Estimation 3.4 Evaluation of Parameter Recovery 4 Results 5 Discussion References Methods to Retrofit and Validate Q-Matrices for Cognitive Diagnostic Modeling 1 Introduction 2 Method 3 Results 4 Discussion and Significance References The Sum Scores and Discretization of Variables Under the Linear Normal One-Factor Model 1 Introduction 1.1 The Linear Normal One-Factor 1.2 Monotone Transformations 2 The LNF Model and the Sum Score 2.1 Preliminaries 2.2 The Sum Score After Discretization of Variables 3 Discussion References Identifying Zones of Targeted Feedback with a HyperbolicCosine Model 1 Introduction 2 Zones of Targeted Feedback 3 Hyperbolic Cosine Model 4 A General Illustrative Example 5 Application to Writing Assessment 6 Discussion References A Reduced Social Relations Model for Dyad-Constant Dependent Variables 1 Introduction 1.1 Standard SRM 1.2 SRM+pred 2 Reduced SRM 3 Motivating Example 3.1 Background 3.2 Method 3.3 Estimating a Reduced SRM 3.4 Results and Discussion 4 General Discussion Appendix References Quality Assurance in Digital-First Assessments 1 Introduction 1.1 Theoretical Framework 1.2 The Context of AQuAA 2 What Statistics Should Be Tracked in the Quality Assurance System? 2.1 Scores 2.2 Test Taker Profile 2.3 Repeaters 2.4 Item Analysis 2.5 Item Exposure 3 How Can the Statistics Be Continuously Monitored and Updated? 4 What Criteria Should Be Used to Identify Irregularities in the Statistics? 5 What Actions Should Be Taken When There Are Alerts of Irregularities? 6 Discussion References Effects of Restoring Missing Data on Uniform Differential Item Functioning Analysis 1 Introduction 2 Method 2.1 Study 1 2.2 Study 2 3 Results 3.1 Study 1 3.2 Study 2 4 Conclusions and Discussion References Two-Step Approach to Topic Modeling to Incorporate Covariate and Outcome 1 Introduction 2 Theoretical Backgrounds 2.1 Structural Topic Model (STM) 2.2 Supervised Latent Dirichlet Allocation (sLDA) Model 3 Methods 3.1 Data Description 3.2 Data Analysis 4 Results 4.1 Study 1: Results of Approach 1, STM and Regression 4.2 Study 2: Results of Approach 2, sLDA and ANOVA 4.3 Comparison of Study 1 and Study 2 5 Conclusion References Modeling Student's Response Time in an Attribute Balanced Cognitive Diagnostic Adaptive Testing 1 Introduction 1.1 CD Models (CDM) 1.2 Item Selection in CD-CAT 1.3 Response Time Framework 2 Methods 2.1 Data 2.2 Evaluation Criteria 3 Results 4 Discussion Appendix References Impact of Construct Reliability on Proposed Measures of Structural Fit When Detecting Group Differences: A Monte Carlo Examination 1 Introduction 2 Latent Variable Modeling 2.1 Model Fit 3 Approaches for Evaluating Structural Fit 3.1 RMSEA-Path 3.2 Conditions 9 and 10 4 Simulation Study 4.1 Method 4.2 Results 5 Discussion References Index