Speakers: Jingchen Liu, Xueying Tang, and Susu Zhang
Date: July 25, 2020 11:00 am – 4:30pm
Format: online via Zoom (click here for pdf of flyer)
Website and registration (free) at http://www.scientifichpc.
Process data refers to log files generated by human-computer interactive items. They contain detailed keystrokes and mouse clicks as well as their timestamps. Our research shows that process data contain substantially more information than classic item responses. They bring great opportunities and at the same time great challenges to the psychometrics community. In this course, we summarize our research methods and empirical findings and provide hands-on training for these methods via real and simulated data. We first provide an overview that features extracted from process data do contain more information than classic item responses. In addition, we also address the question how process data helps solving psychometric problems. To do so, we provide several applications of process data analysis to specific psychometric problems: improving test reliability, reducing/removing differential item functioning, and improving career planning.
We developed an R package ProcData, an open source package for exploratory process data
This course contains three main sections: 1. an overview of process data analyses including an introduction of the methodological development and key empirical results, 2. an introduction to ProcData and hands-on practice, 3. specific applications and hands-on practice of process data to psychometric problems: improving test reliability, reducing/removing differential item functioning, prediction of career development and satisfaction.