Pre-Conference Workshop

The purpose of this workshop is to introduce participants to the use of generalized linear mixed models and Rasch Measurement Theory. A brief overview of generalized linear mixed models is provided. De Boeck and Wilson (2004) described the development of explanatory item response models (EIRM). We illustrate the use of an R package called EIRM (Bulut, 2021) to estimate several extended Rasch models including the Dichotomous Model, the Linear Logistic Rasch Model, Linear Regression Rasch Model, and Combined Covariates Rasch Model.

References: 
Bulut, O. (2021). eirm: Explanatory item response modeling for dichotomous and polytomous item responses [Computer software]. 
Available from http://CRAN.R-project.org/package=eirm. 

De Boeck, P., & Wilson, M. (Eds.) (2004). Explanatory item response models: A generalized linear and nonlinear approach. New York: Springer.  

Prof. Zi Yan, the deputy chair of PROMS, of the Education University of Hong Kong is specialized in educational assessment and applying Rasch analysis in education and psychology issues. This practical workshop will be conducted in Chinese to introduce Rasch measurement to colleagues using Winsteps software. Theoretical and procedural presentations are followed by guided hands-on data analysis and interpretation. Participants were told to bring their own laptop computers (with Office and Adobe Reader pre-installed).

The purpose of this workshop 
is to introduce in both English and Chinese to participants the use of the first Chinese version of Rasch-based Item Analysis and Test Equating System ad hoc developed over the past three years of fighting against COVID-19 under the global situation. We want our participants to bring their own laptop computers because we’ll distribute Rasch-GZ (student version), free of charge, for you to install on your computer simply with a click of mouse. The workshop falls into three parts. During the first two hours, the use of item analysis as reference to moderate multiple-choice questions and test equating via linking items will be introduced. Participants needn’t have particular psychometric competence. As Rasch-GZ displays all the results in Chinese, it would be easy to follow the menu operation. After the coffee break, we’ll demonstrate with real data to show how to run Rasch-GZ. In the afternoon, we’ll have practice with a questions and answers session. 

The background of the study 
Of all the objective measurement approaches at the disposal of researchers and practitioners over the past century, the Rasch model is the most versatile, powerful and feasible. However, in China, such a versatile, powerful and feasible approach is mostly confined within a small professional group of English major for language testing, the understanding, dissemination and the beauty of which are never fully experienced or shared by a majority of other Chinese speaking researchers, especially those young researchers whose majors are not in English language. Furthermore, so far there is practically no testing-specific program producing result interpretations in Chinese. This has been the real status quo ever since the ten-year Matriculation English Test (MET) Equating Project (1990-1999) was implemented in China. Motivated by this, the workshop runner organized a small group of professionals during the COVID-19 pandemic period beginning in 2019, discussing the possibility of meeting the on-going rapid development with an emphasis on further promoting the use of Rasch so as to be well geared with the international practice today.

Aims of the workshop 
Our goal is to present the most updated RASCH-GZ program in Chinese version and publish the user guide handbook that would give the Rasch community of Chinese-speaking researchers and practitioners worldwide a general picture regarding Rasch measurement conducted in the context of PROMS 2023, Macau, China. The past three years witnessed our achievement. We have successfully developed Rasch-GZ and could present and demonstrate with data here.

Features of Rasch-GZ 
To meet the higher demands of test equating today, Rasch-GZ exhibits in at least five aspects as follows:

1. unlimited data matrix;
2. adjustable number of linking items;
3. Chi-square test prior to equating;
4. plotting and online technical supports;
5. language (Chinese or English) can be selected.

Results

The parameters obtained from Rasch-GZ are highly correlated with those from BILOG, PARSCALE and WINSTEPS 

Significance of Rasch-GZ 

1. Rasch-GZ is the first Chinese version of Rasch-based system for item analysis and test equating; 
2. The interface designed has well met the needs of Chinese speaking researchers whose majors are not English;
3. The release of Rasch-GZ will greatly promote the learning and application of Rasch model in and outside China;
4. Student version is free of charge, encouraging our younger researchers to get access to Rasch learning.