Measurement Models: Topics

The table below outlines the topics covered in the three-day measurement models workshop. Measurement models encompass factor analysis models, item response theory models, and some latent class and latent profile mixture models. Measurement models are important for scale development, intermediate outcomes, errors-in-variables models, etc. Briefly, day 1 introduces the kinds of research questions approached using measurement models; an introduction to Mplus software; using Mplus software; and the importance of workflow; finally, measurement models are introduced. Day 2 introduces Item Response Theory (measurement models for ordinal dependent variables) and Differential Item Functioning (DIF) detection. Days 3 & 4 expand the discussion of DIF and introduces special modeling issues and approaches, including models with latent classes, complex sampling weights, and Bayesian data analysis.

Day 1

Introduction

Workshop Objectives
Content Covered in Workshop
Resources to Continue Learning After Workshop
Broad Overview of Measurement-Related Research Questions

Orientation to General Latent Variable Modeling Framework

Path Diagram Notation
Perspectives on Latent Variables

Orientation to Mplus

Writing a Mplus command file
Using Mplus on a PC
Using Mplus on a MAC
Getting Data into Mplus (Using SAS, SPSS, Stata, R)
Workflow and Reproducibility

Factor Analysis

Dimensionality
Model assumptions
- Good Practice
- Bad Practice
Exploratory Factor Analysis
Model Fit Statistics: Interpretation and Use
Confirmatory Factor Analysis (CFA)
Bifactor Analysis
CFA with Covariates
- Multiple Indicator, Multiple Cause Models (MIMIC)
- CFA with Covariates and Direct Effects
- Model Building Strategies (model fitting, decision logic)

Day 2

Item Response Theory (IRT)

Item Response Theory
Test and Item Information
IRT as a general latent variable model
Latent Response Variable Formulation for Categorical Dependent Variables
Assumptions of IRT
Applications of IRT

Day 3

Differential Item Functioning

Uniform Differential Item Functioning
- Detection with CFA with Covariates
- Detection with Multiple Group CFA
- MIMIC Models with categorical dependent variables
Non-uniform DIF
- Multiple Group CFA
- Multiple Group MIMIC
Algorithm for model building with Multiple Group CFA for DIF
- Using Mplus and DIFFTEST procedure to detect uniform DIF
- Automating the DIFFTEST procedure to enhance reproducibility and reduce risk of errors
- Comparison with other methods and software
- Non-Uniform DIF using a single group approach

Day 4

Factor Mixture Models
Incorporating Complex Sampling Weights
CFA with Bayesian Data Analysis techniques
Responsible Conduct of Research
- Confidentiality
- Sensitivity
- Understanding Group Differences Research versus Individual Effects
- Circularity and Ipsativity in Measurement Research
Resources to Continue Learning After Workshop

Google Sites

Report abuse