Members of the RMME Community will share their work in a variety of different research presentations at the 2024 annual meetings of the American Educational Research Association (AERA) and the National Council on Measurement in Education (NCME). Be sure to check out these awesome RMME Community sessions in Philadelphia, PA this month!
Loading...
RMME/STAT Joint Colloquium
In Defense of Unrestricted Spatial Regression
Dr. Dale Zimmerman
University of Iowa
Friday, April 12, at 11AM ET
AUST 202
http://tinyurl.com/rmme-Zimmerman
Spatial regression is commonly used in the environmental, social, and other sciences to study relationships between spatially referenced data and other variables, and to predict variables at locations where they are not observed. Spatial confounding, i.e., collinearity between fixed effects and random effects in a spatial regression model, can adversely affect estimates of the fixed effects, and it has been argued that something ought to be done to "fix" it. Restricted spatial regression methods have been proposed as a remedy for spatial confounding. Such methods replace inference for the fixed effects of the original spatial regression model with inference for those effects under a model in which the random effects are restricted to a subspace orthogonal to the column space of the fixed effects model matrix; thus, they “deconfound” the two types of effects. We prove, however, using classical linear model theory, that frequentist inference for the fixed effects of a deconfounded linear model is generally inferior to that for the fixed effects of the original spatial linear model; in fact, it is even inferior to inference for the corresponding nonspatial model (i.e., inference based on ordinary least squares). We show further that deconfounding also leads to inferior predictive inferences. Based on these results, we argue against the use of restricted spatial regression, in favor of plain old (unrestricted) spatial regression. This is joint work with Jay Ver Hoef of NOAA National Marine Mammal Laboratory and was published in 2022 in The American Statistician.
Loading...
RMME/STAT Joint Colloquium
Some Recent Developments in Educational and Psychological Measurement
Dr. Zhiliang Ying
Columbia University
Friday, March 29, at 11AM ET
In Person: AUST 202
Virtual: https://tinyurl.com/rmme-Ying
Measurement theory plays a foundational role in educational and psychological assessment. Classical item response theory (IRT) models are widely used in the design and analysis of educational tests and psychological surveys that involve multiple choice questions. In this talk, we will first discuss some recent progress related to variations and extensions of the classical IRT model-based methods. We will then turn to the modeling and analysis of process data arising from complex problem-solving items, which are increasingly being adopted in large scale educational assessment. New developments, including statistical models and machine learning algorithms, will be presented. Examples from educational testing and psychological assessment will be used for illustration.
Loading...
Are you traveling to Philadelphia, PA for the Eastern Psychological Association’s 2024 annual meeting? RMME is–and we’d love to meet you! Join us at booth #20 (Franklin Hall A, Downtown Marriott), for an opportunity to chat personally with Dr. Sarah D. Newton, the Associate Director of RMME Online Programs!
What are RMME faculty up to? Check out Dr. Bianca Montrosse-Moorhead’s Evaluation Café talk on February 28, 2024, at 12pm ET, to find out! Her featured presentation is entitled, “Modernizing Evaluation’s Cartography, Architectural Blueprint, and Definition”. Register here to reserve your spot in this excellent session: http://tinyurl.com/EvalCafe2024-BMM
ABSTRACT: In this Evaluation Café presentation, Dr. Montrosse-Moorhead will preview the Country of the Mind map, where the Evaluation Building is located on this map, and a close-up of the building itself. None of these have been visualized before in published scholarship. Dr. Montrosse-Moorhead will also share the proposed amended definition, why it is necessary, and the implications of adopting the amended definitions for evaluation practice; the implications for the instruments, methods, and techniques we use; and the implications for evaluation’s theoretical and metatheoretical scholarship.
The 2024 Modern Modeling Methods (M3) Conference is currently accepting proposals. Review the full call for proposals at: https://modeling.uconn.edu/call-for-proposals-2024/
Submit today! Proposals are due January 29!
For more information, visit: https://modeling.uconn.edu/
RMME/STAT Joint Colloquium
Some New Developments on Pairwise Likelihood Estimation & Testing in Latent Variable Models
Dr. Irini Moustaki
London School of Economics
Friday, December 1, at 11AM ET
https://tinyurl.com/rmme-Moustaki
Pairwise likelihood is a limited-information method used to estimate latent variable models, including factor analyses of categorical data. It avoids evaluating high-dimensional integrals and, thus, is computationally more efficient than full information maximum likelihood. This talk will discuss two new developments in the estimation and testing of latent variable models for binary data under the pairwise likelihood framework. The first development is about estimation and limited information goodness-of-fit test statistics under complex sampling. The performance of the estimation and the proposed test statistics under simple random sampling and unequal probability sampling is evaluated using simulated data. The second development focuses on computational aspects of pairwise likelihood. Despite its computational advantages it can still be demanding for large-scale problems that involve many observed variables. We propose an approximation of the pairwise likelihood estimator, derived from an optimization procedure relying on stochastic gradients. The stochastic gradients are constructed by subsampling the pairwise log-likelihood contributions, for which the subsampling scheme controls the per-iteration computational complexity. The stochastic estimator is shown to be asymptotically equivalent to the pairwise likelihood one. However, finite sample performances can be improved by compounding the sampling variability of the data with the uncertainty introduced by the subsampling scheme. We demonstrate the performance of the proposed method using simulation studies and two real data applications.
Loading...
RMME/STAT Joint Colloquium
Hidden Markov Models with an Unknown Number of Hidden States
Dr. Xinyuan Song
The Chinese University of Hong Kong
Friday, November 3, at 10AM ET
Hidden Markov models (HMMs) are valuable tools for analyzing longitudinal data due to their capability to describe dynamic heterogeneity. Conventional HMMs typically assume that the number of hidden states (i.e., the order of HMMs) is known or predetermined through criterion-based methods. This talk discusses double-penalized procedures for simultaneous order selection and parameter estimation for homogeneous and heterogeneous HMMs. We develop novel computing algorithms to address the challenges of updating the order. Furthermore, we establish the consistency of order and parameter estimators. Simulation studies show that the proposed procedures considerably outperform the commonly used criterion-based methods. An application to the Alzheimer’s Disease Neuroimaging Initiative study further confirms the utility of the proposed method.
Loading...
RMME/STAT Joint Colloquium
Uncovering the Hidden Complexity of Statistical Models
Dr. Wes Bonifay
University of Missouri
Friday, October 13, at 11AM ET
https://tinyurl.com/rmme-Bonifay
Model complexity is the ability of a statistical model to fit a wide range of data patterns. Complexity is routinely assessed by simply counting the number of freely estimated parameters in a given model. However, complexity is also affected by configural form, that is, by the particular arrangement of the variables in the model. Recent considerations of configural complexity have found that certain models have an inherent tendency to fit well to any possible data (sometimes achieving superior goodness-of-fit when compared to alternative models that contain a greater number of free parameters!). In this talk, Dr. Bonifay will present a method for evaluating configural complexity and demonstrate how more sophisticated considerations of complexity can improve applied research in the social sciences.
Loading...
Register today for the 2023 Modern Modeling Methods Conference. Conference registration ends June 19!
2023 Modern Modeling Methods (M3) Conference
Dates: June 26 – June 28, 2023
Location: University of Connecticut’s Main Campus in Storrs, CT
Description: The Modern Modeling Methods (M3) Conference is an interdisciplinary conference designed to showcase the latest modeling methods and to present research related to these methodologies. Planned events include:
See the M3 Preliminary Program for a full list of talks.
Visit our website: modeling.uconn.edu.
Reload document 
Open in new tab