The Quantitative Psychology colloquium series meets weekly in the Autumn and Spring semesters. The primary activity is presentation of ongoing research by students and faculty of the quantitative psychology program. Often, guest speakers from allied disciplines (e.g., Education, Statistics, and Linguistics) within and external to The Ohio State University present on contemporary quantitative methodologies. Additionally, discussions of quantitative issues in research and recently published articles are often conducted during colloquium meetings.
Faculty coordinator: Dr. Gyeongcheol Cho
Venue: Psychology 35 and online
Time: 12:30-1:30pm
Please contact Dr. Gyeongcheol Cho if you would like information on how to attend these events over zoom.
August 28, 2023
Organizational Meeting
September 4, 2023
Labor Day
September 11 , 2023
Speaker: Dr. Xiaoming Zhai
Department of Mathematics, Science, and Social Studies Education; University of Georgia
*joint event with University of Maryland, College Park; University of North Carolina at Chapel Hill, University of Notre Dame, Vanderbilt University; University of South Carolina; and University of Virginia
Title: AI and Formative Assessment: The Train Has Left the Station
Abstract: Researchers have been questioning AI –whether we can and should use AI for formative assessment. AI is already being employed, for better or worse, to facilitate formative assessment in various educational contexts. In this talk, Dr. Zhai will demonstrate research on AI-based assessment in science education. He will respond to the many concerns raised in the field regarding AI being used for formative assessment practices. He argues that the focus of research should be on how we can help educators manage an AI revolution that has outpaced a research community caught unaware and recognize the remarkable progress that AI has contributed to formative assessment.
September 18, 2023
Speaker: Dr. Gyeongcheol Cho
Title: Exploring Structured Factor Analysis (SFA): From Theory to Practice – Part I (Theory)
Abstract: Jöreskog’s Covariance Structure Analysis (CSA; 1978), along with its maximum likelihood (ML) estimator, has long stood as the standard approach to structural equation modeling (SEM). Despite its wide adoption, it suffers from two persistent limitations: the occurrence of improper solutions (e.g., negative variance estimates) and the lack of an internal tool for making probabilistic inferences on true factor scores (e.g., estimating the probability that two individuals have different true factor scores). Addressing these challenges, Cho & Hwang (2023) proposed Structured Factor Analysis (SFA), a novel data matrix-based approach to SEM. SFA concurrently estimates model parameters and factor score distribution from a given data matrix, which prevents improper solutions and facilitates the probabilistic inference of true factor scores.
This workshop is designed to offer a succinct review of SFA theories and to grant participants the opportunity for hands-on experience with SFA through its associated software, SFA Prime (https://sfaprime.com/). Attendees are encouraged to bring their personal laptops with the software pre-installed. The SFA Prime software is freely accessible on Windows 10/11 and Mac OS 11.6/12 platforms. Additionally, installing R is recommended for individuals interested in conducting further analyses.
September 25, 2023
Title: Paper discussion on Widaman, K. F. (2018). On common factor and principal component representations of data: Implications for theory and for confirmatory replications. Structural Equation Modeling: A Multidisciplinary Journal, 25(6), 829–847.
Discussant: Dr. Gyeongcheol Cho
October 2, 2023
Symposium on Constructs and Their Measurement
*zoom talk
Speakers: Jolynn Pek & Paul De Boeck, The Ohio State University
Title: Reconsidering the Uncertain Nature of Constructs
Abstract: Constructs are fundamental to psychological research, and their highly abstract nature could be a source of uncertainty in research findings. To address the uncertain nature of constructs, researchers have either (a) moved away from constructs toward specific observables and effects or (b) moved toward constructs with sharper definitions and stricter measurement practices. We propose an alternative formulation of constructs that acknowledges their uncertain nature by recognizing that constructs (a) as composite in nature to allow for heterogenous content (cf. unitary), (b) might be organized in hierarchical systems with overlap, and (c) are variable in nature as reflected in variable measurements. Constructs can thus be visualized as areas within a space instead of a point in which measures might not fully cover their content, pointing to the need for multiscale approaches to measurement. This reformulation seems more consistent with observed variability of findings, which aligns with ongoing and newer methodological avenues that emphasize generalization efforts while preserving and facilitating the integrative and explanatory role of constructs.
Speaker: Rick Hoyle, Duke University
Title: Tightening the Connection between the Delineation and Measurement of Constructs
Abstract: Rigorous quantitative research in psychology requires effective measurement of well-delineated constructs. Poorly specified constructs do not provide adequate information for developing and evaluating measures. And measures that do not fully capture constructs when they are well-specified do not allow for clear inferences that inform theory. In short, rigorous and informative psychological science requires tight connections between well-delineated constructs and high-quality measures of them. I begin with a discussion of types of constructs typical of psychological theories and conceptualizations. I then present approaches to measurement available for operationally defining psychological constructs. Finally, I present an example of a new measure of political sectarianism that features a clear and detailed delineation of the construct, a thoughtfully designed measure of it, and a tight connection between them. I conclude with a discussion of ways the construct-measure connection can be tightened to increase the rigor and informativeness of quantitative research in psychology.
Speakers: Yang Liu & Jolynn Pek, University of Maryland & The Ohio State University
Title: Reliability Calculations in Nonlinear Measurement Models
Abstract: In the light of measuring constructs, reliability is recognized as an important quantification. Given a latent variable measurement model that is assumed to map onto constructs of interest, reliability coefficients reflect how closely observed scores are aligned with latent scores (about the construct). When a nonlinear measurement model is employed, reliability can be computed in different ways, having distinct implications for substantive research. To begin, we explain how different definitions of reliability can be applied to nonlinear measurement models and why they lead to different reliability coefficients. An empirical illustration is provided, which is based on Magnus and Liu’s (2022) analysis of the depressive symptom survey data from the National Comorbidity Survey Replication (NCS-R). A two-dimensional hurdle graded response model was fitted to the data, and reliability coefficients were computed for various pairs of observed and latent scores. The presentation is concluded with a discussion on the usefulness of reliability coefficients (in relation to measuring constructs), or lack thereof, from a practical standpoint.
October 9, 2023
Speaker: Dr. Beth Tipton
Department of Statistics and Data Science; Northwestern University
*joint event with University of Maryland, College Park; University of North Carolina at Chapel Hill, University of Notre Dame, Vanderbilt University; University of South Carolina; and University of Virginia
Title: Generalizability and heterogeneity: Designing your study when treatment effects vary
Abstract: Randomized trials are now common in medicine, the social sciences, and in policy-related research. The primary goal of a randomized trial is to estimate the average treatment effect, which averages over unit-specific treatment impacts that might be quite heterogenous. Often the results of such a trial are to be used for making policy or practice decisions in a target population; this, paired with the fact that most such trials take place in samples of convenience, has led to an increased interest in methods for generalizing causal effects. This has included new methods for estimation of average treatment effects that take into account population data, as well as methods for designing sampling and recruitment strategies for such studies. But if treatment effects actually vary, often researchers seek to understand if such heterogeneity can be predicted or explained by unit characteristics. In this talk, I focus on this sampling problem in the face of heterogeneity. I show that methods for optimal designs found in response surface models can be useful, too, in designing sampling plans that result in increased power and precision for these moderator effects. I situate this in an example based on an evaluation of a school-based reading program.
October 16, 2023|
Speaker: Dr. Gyeongcheol Cho
Title: Exploring Structured Factor Analysis (SFA): From Theory to Practice – Part II (Practice)
Abstract: Jöreskog’s Covariance Structure Analysis (CSA; 1978), along with its maximum likelihood (ML) estimator, has long stood as the standard approach to structural equation modeling (SEM). Despite its wide adoption, it suffers from two persistent limitations: the occurrence of improper solutions (e.g., negative variance estimates) and the lack of an internal tool for making probabilistic inferences on true factor scores (e.g., estimating the probability that two individuals have different true factor scores). Addressing these challenges, Cho & Hwang (2023) proposed Structured Factor Analysis (SFA), a novel data matrix-based approach to SEM. SFA concurrently estimates model parameters and factor score distribution from a given data matrix, which prevents improper solutions and facilitates the probabilistic inference of true factor scores.
This workshop is designed to offer a succinct review of SFA theories and to grant participants the opportunity for hands-on experience with SFA through its associated software, SFA Prime (https://sfaprime.com/). Attendees are encouraged to bring their personal laptops with the software pre-installed. The SFA Prime software is freely accessible on Windows 10/11 and Mac OS 11.6/12 platforms. Additionally, installing R is recommended for individuals interested in conducting further analyses.
October 23, 2023
Speaker: Dr. Yoonjung Yoonie Joo
Title: Quantitative Methodologies in Computational Genetics: A Path to Personalized Health
Abstract: A massive number of population-based databases have become available recently, providing novel research opportunities for the interrogation of DNA genotype-phenotype associations on unexplored clinical landscapes. The extensive phenotypic information encoded in large-scale EHR (electronic health records) databases, ranging from diagnosis code, physician reports, brain neuroimaging data to DNA genotype data, are valuable resources for clinical researchers to characterize the pleiotropic architecture of human complex traits. My research mission is to establish novel and scalable data-driven frameworks to efficiently utilize patient genetic & clinical data for developing various risk prediction models for the goal of precision medicine, in combination with recent advanced computational methodologies, including natural language processing(NLP), supervised/unsupervised machine learning(ML), and deep learning techniques.
In this seminar, I will provide a comprehensive overview of the application of several methodologies in risk prediction and causal modeling of health outcomes from a genomic perspective. To be specific, I will present the following scientific questions addressed by large-scale DNA biobank data with several pivotal data-driven methodologies: (1) identification of individuals at high risk for polycystic ovary syndrome(PCOS) / suicidal behaviors with their genetic and diagnosis data: a polygenic and phenotype risk score (PPRS) prediction model approach in multi-ancestry participant; (2) utilization of genetic data for predicting rate of cognitive decline in later ages with phenome-wide association studies; (3) investigation of causal effects of late/early menopause ages on various health outcomes with Mendelian Randomization (MR), a method of using measured variation in genes of known function to examine the causal effect of a modifiable exposure on disease in observational studies, etc. This seminar is particularly relevant for researchers interested in the transformative potential of genomic data science in advancing our understanding of the social sciences.
October 30, 2023
Speaker: Dr. Alexander Wasserman
Title: Applications of Developmental Science to High-Risk Youth: A Risk and Resilience Perspective
Abstract: Adolescence is a time of significant development, such as changes in relationships with parents and peers, pubertal maturation, and changes in brain structure and function. Co-occurring with these normal changes, adolescence is also a developmental stage when the propensity to take unnecessary risks is especially high. My research program aims to elucidate developmental mechanisms through which these developmental changes are related to risk behavior (e.g., substance use, offending). Importantly, for high-risk youth, including those with a family history of substance use disorder or those involved with the justice system, the transition to adolescence may be especially challenging. During my presentation, I present two studies of high-risk youth. In the first study, I examine if youth with a family history of substance use disorder are more likely to engage in heavy substance use compared to low-risk youth because of differences in the development of impulse control (e.g., planning ahead) and sensation seeking (e.g., thrill-seeking behaviors like riding a roller coaster). In the second study, I examine if the influence of impulse control and sensation seeking on offending is stronger or weaker during adolescence compared to adulthood in a sample of justice-involved youth.
In addition to my research program focused on high-risk youth, I have extended my research to Latinx youth development. Latinx youth are at elevated risk of internalizing problems. Taking a cultural resilience perspective, in a third study, I examine if parent or peer support can buffer against the deleterious effect of bicultural stressors (e.g., pressure to learn the “American” ways of doing things) on Latinx youth’s internalizing problems.
Robert Wherry Speaker
November 2, 2023
Thursday 3:00-4:00pm (zoom)
Speaker: Dr. David J. Hand
Imperial College London
Title: Size matters: Measurement and understanding
Abstract: Measurements are so ubiquitous that we often fail to notice them: they are simply parts of the conceptual universe in which we function. However, there’s more to measurement than meets the eye, and misunderstanding measurement can lead to mistakes and even disasters. I examine just what constitutes measurement, how measures are constructed, how to interpret them, and how to use them. The talk is illustrated with measurement tools from psychology, medicine, physics, economics and other areas.
November 6, 2023
Speaker: Dr. Mijke Rhemtulla
Department of Psychology, University of California, Davis
*joint event with University of Maryland, College Park; University of North Carolina at Chapel Hill, University of Notre Dame, Vanderbilt University; University of South Carolina; and University of Virginia
Title: Consequences of Mistaking the Measurement Model in SEM, Alternatives to Common Factors, and a Method for Model Selection
Abstract: Much of the appeal of structural equation models lies in their capacity to account for measurement error by modeling abstract constructs (extraversion, quality of life, school readiness) as latent common factors. This feature of SEMs has led researchers across the social sciences to use latent variable SEMs with little awareness of the assumptions that the reflective measurement model requires. But the choice of measurement model carries implications about the structure of data and about data-construct associations: An incorrect model can change the meaning of the construct and render structural relations uninterpretable. When a common factor model is mis-applied, structural model coefficients can be (highly) biased, and this bias can arise even when model fit is perfect. Recent developments allow for two alternative measurement models to be implemented in SEM: a composite score model with user-defined weights, and a composite score model with model-estimated weights. In this talk I discuss the problem and present the alternative measurement models, and I propose a statistical test that may help to identify the most appropriate measurement model given data.
November 13, 2023
Speaker: Frank Leyva Castro
Title: Evaluating Assumptions in Structural Equation Modeling: An Extension of Diagnostic Approaches
Abstract: Structural Equation Modeling (SEM) is a powerful statistical tool that allows for the analysis of complex linear relationships among variables. Because of this SEM remains as one of the most popular statistical tools to analyze psychological constructs. Despite its widespread use, there is a literature gap on how to effectively evaluate the assumptions underlying SEM. This presentation, based on my thesis work, shows my work to bridge this gap by extending diagnostic practices from multiple linear regression, factor analysis, and multilevel models to SEM. By reviewing the similes between multilevel modeling and SEM, we bring graphical diagnostics from regression analysis. These extended approaches can aid psychology researchers in evaluating the tenability of their SEM results. The expected outcome is a clear procedure on how to evaluate assumptions when using SEM, encouraging good practices in psychological research.
November 20, 2023
Speaker: Shannon Jacoby
Title: Assessing Monotonicity and Trends: A Simulation Study
Abstract: Across a wide array of substantive research areas within psychology, many researchers make use of the regression modeling framework. A central assumption of this framework is that the relationship between the IV and expected value of the DV is linear, and the current diagnostic criterion for this assumption is a visual inspection of a scatter- or residual plot for the absence of systematic nonlinearity. Given that the violation of the linearity assumption can lead to biased parameter estimates and error variances, the current work endeavors to create an additional avenue for examination of the linearity assumption that goes beyond graphical analysis. Borrowing inspiration from the ANOVA framework, contrast analyses were applied to simulated data as a way to detect linear and quadratic trends in continuous data as well as nonmonotonicity via an ordinal analysis approach. As not all monotonic relations are linear and not all quadratic trends are violations of monotonicity, this work investigates both trends and violations of monotonicity in an effort to more fully understand IV/DV relationships under a variety of conditions. Type I errors, power, and comparison of findings with additional methods are discussed, as well as limitations and directions for future research.
November 27, 2023
Speaker: Dr. Melanie Wall
Department of Biostatistics, Columbia University
*joint event with University of Maryland, College Park; University of North Carolina at Chapel Hill, University of Notre Dame, Vanderbilt University; University of South Carolina; and University of Virginia
Title: Incorporating intersectionality using latent class analysis within health contexts
Abstract: Intersectionality posits that social categories (e.g. race, gender, sexual orientation) and the forms of social stratification that maintain them (e.g. racism, sexism, homophobia) are interlocking, not discrete. An intersectionality framework considers harms and oppression and also privileges and unearned advantages. By focusing on intersectionality, we can examine axes of social power that underlie our overall health and the systems that support it with the goal of identifying levers for change. A recent systematic review (Bauer et al 2022 Social Psychiatry and Psych Epi) demonstrated a growing use of latent variable methods including latent class analysis for applications of intersectionality. Latent class analysis (LCA) has been described as a “person-centered” approach as it clusters within-individual characteristics seen to be appropriate to intersectionality. In the present talk, I will demonstrate the use of LCA for combining intersecting social positions with multiple factors characterizing an initial mental health encounter. The example comes from a study of ethnoracial disparities in coordinated specialty care for people with psychosis. Clusters were identified based on the first-contact experience (i.e., referral source, type of first mental health service contact, symptoms at referral) in combination with sociodemographic variables impacting an individual’s social position (age, gender, ethnoracial group, language proficiency, sexual orientation, living situation, type of insurance, homelessness, and urbanicity). Visualizations of intersectional cluster results and comparisons between the LCA approach and analyses focused on each variable separately will be presented.
December 4, 2023
Brief presentations on external talks by the following students:
Frank Leyva Castro
Shannon Jacoby
December 7 ,2023 (Thursday)
Venue & Time: Lazenby 120, 2:30-4:00pm
*joint event with the Social Psychology area
Speaker: Dr. Rick H. Hoyle
Department of Psychology and Neuroscience, Duke University
Title: Intellectual Humility and the Management of Personal Knowledge and Beliefs in an Uncertain and Polarized World
Abstract: Differing assumptions, experiences, and motives contribute to multiple perspectives on and conclusions about what is right and true. The result is alternative views concerning nearly all topics and issues and the potential for animosity and conflict when people are exposed to views that differ from their own views. That potential is substantially reduced for people who accept that their views could be wrong or ill-informed and might benefit from thoughtful consideration of new information. More generally they recognize and accept the possibility that their views may be less defensible than views promoted by other people and information sources. And, even if they remain convinced that a current view does not warrant revision or change, they do not view challenges to it as a personal affront. That is to say, they react to alternative views with intellectual humility. In this talk, I provide an overview of a program of research focused on the conceptualization and measurement of intellectual humility with an emphasis on psychological mechanisms that may serve as targets of interventions that aim to promote intellectual humility. I hope to stimulate discussion of the intellectual humility construct understood both as an individual difference and as a framework for productive conversation about divisive issues.