Kang Xue, PhD
Psychometrician
Kang Xue works with the MAP Reading Fluency assessment. His research interest centers on the intersection of psychometrics, machine learning, data science, and artificial intelligence. His interdisciplinary background allows him to borrow some ideas from machine learning and apply them to psychometric models, such as IRT and CDM, with the hope of better handling complex assessment data. Dr. Xue joined NWEA in 2020 after obtaining a PhD in educational psychology from the University of Georgia. Dr. Xue holds another doctoral degree in electrical engineering, and he served as a senior research engineer at Samsung Advanced Institute of Technology (Advanced Technology Lab) from 2013 to 2015.
Publications by Kang Xue
English MAP Reading Fluency technical report
This technical report documents the processes and procedures employed by NWEA to build and support the English MAP Reading Fluency assessment.
Products: MAP Reading Fluency
Topics: Computer adaptive testing, Reading & language arts, Test design
This study examines the text quality of math assessment items for students with VI who use screen readers. Using data from about 29.5 million students taking standard versions of the MAP Growth math assessment, and 48,845 students taking accessible versions, we identified high-quality items, those that measured achievement for both students with and without VI equally well, and low-quality items, which showed differences between the two groups of students.
Topics: Accessibility, Equity, Math & STEM
A semi-supervised learning-based diagnostic classification method using artificial neural networks
This research study is the first time of applying the thinking of semi-supervised learning into CDM. Also, we used the validating test to choose the appropriate parameters for the ANNs instead of using typical statistical criteria, such as AIC, BIC.
Topics: Measurement & scaling
In data collected from virtual learning environments (VLEs), item response theory (IRT) models can be used to guide the ongoing measurement of student ability.
Topics: Measurement & scaling
Predicting the difficulty and response time of multiple choice questions using transfer learning
This paper reports on whether transfer learning can improve the prediction of the difficulty and response time parameters for ≈ 18,000 multiple-choice questions from a high-stakes medical exam.
