Research questions regarding how, for whom, and where a treatment achieves its effect on an outcome have become increasingly valued. Such questions can be answered by causal moderated mediation analysis, which assesses the heterogeneity of the mediation mechanism underlying the treatment effect across individual and contextual characteristics. The purpose of this three-hour virtual course is to introduce the general definition, identification, estimation, and sensitivity analysis for causal moderated mediation effects under the potential outcomes framework. Participants will also learn how to use a user-friendly R package to conduct the analysis and visualize analysis results. The method introduction and the package implementation will be illustrated with a re-analysis of the National Evaluation of Welfare-to-Work Strategies (NEWWS) Riverside data.
GSCA-SEM (generalized structured component analysis structural equation modelling) is an umbrella term that includes three SEM methods—GSCA, GSCA M , and IGSCA—for estimating models with components only, with factors only, or with both factors and components, respectively. GSCA-SEM is highly versatile in accommodating the two statistical representations of constructs. GSCA Pro is a stand-alone software program for GSCA-SEM. The software can be freely downloaded from its website (www.gscapro.com). It provides a graphical user interface that allows users to draw their model as a path diagram easily, fit GSCA-SEM to the model, and obtain results.
This workshop begins by explaining the conceptual foundations of GSCA-SEM, focusing on model specification and evaluation. It then provides step-by-step illustrations of using the free software for various GSCA-SEM applications.
The workshop will cover meta-analytic structural equation modeling (MASEM), which uses the techniques of meta-analysis and structural equation modeling to synthesize correlation matrices and fit hypothesized models on the combined correlation matrix. It can be used to test path models, confirmatory factor analytic models, and structural equation models from a pool of correlation matrices. MASEM offers the benefits of both meta-analysis and SEM.
During the workshop, I will provide an introduction to the basic theory of MASEM and demonstrate how to conduct the analyses with R. While some familiarity with R would be beneficial, the workshop is designed to be accessible to those who are new to the programming language.
The International Society for Data Science and Analytics-ISDSA Nigeria will organize a series of Data Science Training in 2024.
Modeling longitudinal data is one of the most active areas of research in social, behavioral, and education sciences because longitudinal data can provide valuable insights into change and causal relationships. The application of Bayesian methods in longitudinal research has gained increasing popularity. This interactive workshop focuses on Bayesian methods in analyzing longitudinal data. Particularly, it will cover topics on growth mixture modeling, missing data analysis, and Bayesian model assessment. Concrete examples will be provided to illustrate how to compute, report, and interpret Bayesian modeling results with empirical psychological data.
Dr. Tong is an associate professor at the University of Virginia. Her research focuses on developing and applying statistical methods in the areas of developmental and health studies. Methodologically, she is interested in Bayesian methodology, growth curve modeling, and robust structural equation modeling with nonnormal and missing data. Substantively, she is interested in analyzing the longitudinal development of cognitive ability and achievement skills.
This workshop is supported by the William K. and Katherine W. Estes Fund that is jointly overseen by the Association for Psychological Science and the Psychonomics Society.
Power analysis is a statistical method used in psychological research to determine the appropriate sample size for a study and is often required in grant proposals and by journals as part of the research submission process. WebPower is a software platform that provides a user-friendly interface for conducting power analyses for various statistical methods, including t-tests, ANOVA, regression, mediation analysis, multilevel modeling, longitudinal data analysis, and structural equation modeling. The workshop will focus on teaching participants how to use this software to conduct power analyses for their own research. The workshop was taught at the 2023 APS Annual Convention. You can find the slides and code used by the workshop here.
Latent class modeling (LCM), a branch of latent variable modeling, has become increasingly popular among social researchers in recent years because of its superior capability for detecting unobserved heterogeneity in data. Unlike the traditional latent variable models that focus on extracting factors, the latent variable in a LCM is discrete and categorical，and its groups, called latent classes, provide a classification structure or statistical taxonomy for recovering the sub-population behind the sample. This workshop will briefly introduce the methodological concepts of LCM and pay more attention to the analytic techniques, using Mplus, for implementing latent class analysis as well as latent profile analysis along with a major extension for longitudinal data analysis, entitled latent transition analysis (LTA), which is used to examine how individuals transition in the latent class membership over time.
Dr. Hawjeng Chiou is a distinguished professor of the College of Management at National Taiwan Normal University. His major interest is in applied psychometrics, particularly for the applications of advanced modeling techniques, such as Structural Equation Modeling, Multilevel Linear Modeling, and Latent Class Modeling, with substantial areas such as Human Resource Management, Organizational Behavior, and Psychological Testing as well as Creativity research. He has published multiple books on a variety of topics.
In social sciences, an input variable often affects an outcome variable via a third variable, or mediator. Mediation analysis is used to answer the question of “how”/ “why” the input variable affects the outcome. This workshop will introduce the principles and methods (e.g., multivariate regression, path analysis, multilevel models, etc.) for mediation analyses. Emphasis will be on data analysis and interpretation. After the workshop, attendants should be able to propose research questions related to mediation analyses, use advanced methods such as bootstrap to test mediation effects, decide which type of model is appropriate for mediation, interpret and display empirical findings, and understand some other important issues of mediation research such as longitudinal designs for assessing mediation and missing data problems.
Dr. Laura Lu is an associate professor in the Quantitative Methodologies (QM) program at the University of Georgia (UGA). In general, she has expertise in structural equation modeling (SEM), longitudinal data analysis, hierarchical linear modeling (HLM), and computational statistics such as Bayesian. During her professional career at UGA, her research has focused on developing innovative statistical approaches to address perennial challenges in statistical modeling such as mixture structure, reliabilities, model selection, missing data, outliers, topic modeling, and also promoting statistical models to applied research areas through collaborating, mentoring, and classroom teaching.
Dr. Qian (Jackie) Zhang is an associate professor in the Department of Educational Psychology and Learning Systems at Florida State University. Her research interests focus on causal inferences, longitudinal data analysis, and effect size synthesis using multilevel models and multilevel structural equation models. She is also interested in handling sub-optimal data conditions with missing data, measurement error, and/or confounding variables in statistical analyses. She collaborates with substantive researchers in the areas of early childhood development, sports psychology, and educational research. Her work has been published in prestigious journals such as Psychological Methods, Multivariate Behavioral Research, Structural Equation Modeling, British Journal of Mathematical and Statistical Psychology, and Behavioral Research Methods.
Longitudinal Data Analysis (LDA) is a popular statistical method in various fields, including psychology, education, sociology, management, economics, and medicine. This workshop aims to provide participants with an overview of the applications and capabilities of LDA. Topics include: (a) introduction of longitudinal designs and longitudinal data, (b) latent growth curve models and growth mixed models, (c) cross-lagged panel models, (d) separating within- and between-person effects models, and (e) advances in LDA. The workshop will consist of lectures and software demonstrations using Mplus and R with practical examples.
Dr. Hongyun Liu is a professor in Faculty of Psychology at Beijing Normal University and has been working on the research of quantitative psychology for nearly 20 years. Her main research areas are psychological statistics, the theory and applications of psychological and educational measurement, and quantitative research methods. Her research interests involve the Internet-based test development, the theory and application of educational and psychological assessment, data analysis methods in large-scale assessments, and longitudinal data analysis.
Social network analysis is becoming increasingly popular in social, educational, and psychological sciences. This interactive course intends to provide participants with a detailed introduction, practical examples, and demonstration of analyzing social network data using the free software R. Topics covered include (1) Network Data; (2) Network Visualization; (3) Network Statistics; (4) Basic and Advanced Network Models. Especially, we will cover classical models such as Stochastic Block Model, Exponential Random Graph Model, Latent Space Model, and the newly developed techniques such as Latent Factor Space Modeling and Network Mediation Analysis.
Dr. Haiyan Liu is currently an assistant professor of Quantitative Methods, Measurement, and Statistics at the University of California, Merced. Dr. Liu’s research centers broadly on statistical modeling of psychological and behavioral data such as high-dimensional data, longitudinal data, and categorical data. Her recent research includes social network modeling, Bayesian structural equation modeling, and non-parametric modeling of growth curves.
Deep learning is a very active area of research in the machine learning and artificial intelligence communities. The foundation of deep learning is the neural network. In this workshop, we will teach how to practically carry out deep learning using the R package keras. Several real-world data sets will be used to illustrate how to build single layer, multiple layer, convolutional, and recurrent neural networks. Basic knowledge of R is expected.
Instructor: Prof. Zhiyong Zhang.
Dr. Zhiyong Zhang is a professor in Quantitative Psychology at the University of Notre Dame. His research aims to develop better statistical methods and software in the areas of education, health, management and psychology. He has conducted research in the areas of Bayesian methods, Big data analysis, Structural equation modeling, Longitudinal data analysis, Mediation analysis, and Statistical computing and programming. His most recent research involves the development of new methods for social network and text analysis.
The class is free. Please register early.
This workshop is designed to introduce the fundamentals and practical applications of modeling data using data mining and machine learning methods. It surveys a series of topics starting from exploratory data analysis, unsupervised learning, supervised learning and other learning methods to uncover the patterns and structure of data targeting practical problem solving and identifying solutions. Students will learn by hands-on programming including such methods as regression, classification, logistic models, non-linear models, tree-based models, association rules, neural network and support vector machines. New developments in learning such as model visualization, deep learning and interpretable machine learning will also be introduced and illustrated.
Social network analysis is becoming increasingly popular in social, educational, and psychological sciences. This interactive course intends to provide participants with a detailed introduction, practical examples, and demonstration of analyzing social network data using the free software R. Topics covered include (1) Network Data; (2) Network Visualization; (3) Network Statistics; (4) Basic and Advanced Network Models. Especially, we will cover classical models such as Stochastic Block Model, Exponential Random Graph Model, Latent Space Model, and the newly developed techniques such as Latent Factor Space Modeling and Network Mediation Analysis.
Modeling longitudinal data is one of the most active areas of research in social and behavioral sciences because longitudinal research provides valuable insights into change and causal relationships. The application of Bayesian methods in longitudinal research has gained increasing popularity. Taught by four quantitative researchers who are active in developing Bayesian methods and longitudinal data analytical methods, this workshop will teach participants how to analyze longitudinal data using Bayesian statistics. We will introduce the basic idea of Bayes' theorem first and move on to models including multilevel models, growth curve models, growth mixture models, and longitudinal structural equation models. Concrete examples will be provided to illustrate how to compute, report, and interpret Bayesian modeling results with empirical psychological data. Additionally, we will teach the application of Bayesian robust methods for nonnormal data ignorable and non-ignorable missing data.
This workshop teaches researchers how to conduct a statistical power analysis to determine the sample size for a planned study when applying structural equation modeling. Participants will learn statistical power analysis for mediation analysis, path analysis, and structural equation modeling using both traditional methods and Monte Carlo methods. Free software WebPower (https://webpower.psychstat.org) will be used to illustrate how to conduct power analysis in practice.
- Understanding Machine Learning Concepts
- Data Preprocessing and Feature Engineering
- Model Selection and Evaluation
- Implementing Machine Learning Algorithms
- Hyperparameter Tuning
- Model Deployment and Integration
- Ethical Considerations and Bias in Machine Learning
- Continuous Learning and Model Maintenance
You'll learn the fundamentals of data analysis with Python. Through the workshop, you will understand the six steps of data analysis processes as well as know how to read data from sources like CSV files and SQL, and how to use libraries like Numpy, Pandas, Matplotlib, and Seaborn to process and visualize data.
This Training is design to introduce the foundamental on how to shape and transform your data before the data analysis using Power Query Editor. You will discover how to navigate this intuitive tool and get to grips with Power BI’s Data, Model, and Report views. You’ll load multiple datasets in the Data view, build a data model to understand the relationships between your tables in Model view, and create your first bar graph and interactive map visualization in Report view. You’ll also practice using Power Query Editor to prep your data for analysis.
Through hands-on exercises, you’ll learn how to change and format a wide range of visualizations, before moving on to sorting data and creating hierarchies—making it possible for you to drill into your reports.
