Investigating the psychometric properties of english vocabulary size test using network data analysis

Network science as a new field in psychological measurement provides a good platform for test analysis, but in Iran, research on the analysis of test questions using network data analysis has been neglected.

The aim of this study was to introduce network data analysis as a psychometric-mathematical technique and its use in the questionnaire analysis process.

This research is descriptive-analytical, the statistical population of all language learners was the top ten language teaching institutes in Tehran in 2020, from which 1556 language learners were selected by convenience sampling method and answering the English vocabulary size test (Nation and Waring, 1997).  Data analysis was performed by qgraph software package under R software.

The graph obtained from the communication network of questions based on the Fruchterman-Ringold algorithm was plotted, the parent-child relationship of the questions was determined based on the Bayesian network, and the descriptive features of the test and the indicators of network centrality were extracted.

Introducing the technique of analyzing network data and using it for  questions analysis process is one of the achievements of this article. Using the advantages of network science such as the ability to simplify, manage, store and collect high-volume data and analyze test questions based on the map and indicators that are fully consistent with it, we can understand the internal relationship between different test questions, and by considering the importance of the questions in the network structure of the net relationship between the questions, it is possible to improve the testing methods. Similar studies and researches are proposed to compare the results of network data analysis with psychometric properties obtained based on conventional and classical methods.