Scholarlink

<p>In cognitive diagnostic tests, examinees may adopt different strategies for answering items, such as solution behavior and rapid guessing behavior. In high-stakes tests, examinees may resort to rapid guessing for difficult and complicated items due to time pressure. In addition, examinees with low motivation may constantly switch between solution behavior and rapid guessing behavior to complete the test. Solution behavior signifies thoughtful consideration, while rapid guessing suggests minimal effort. This paper proposes a mixture hierarchical model that utilizes responses and response times to detect rapid guessing behavior in cognitive diagnosis assessments. The model enables the determination of whether examinees have mastered the required knowledge attributes. A Markov chain Monte Carlo method is introduced for estimating the parameters of the model. Simulation studies show that the parameter estimates obtained using this new model are more accurate than those obtained using a non-mixture model under various conditions. Bayesian assessment criteria are employed to evaluate the model fitting. Finally, a practical application is presented to illustrate the utility of the proposed model.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Advancing classification methods in computerized testing: a novel mixture hierarchical cognitive diagnosis model incorporating response times for assessing rapid guessing behavior

  • 期刊论文

  • 发表日期：

    2025年2月18日

查看全文
  • Lidan Liang,
  • Jiwei Zhang,
  • Jing Lu

摘要

In cognitive diagnostic tests, examinees may adopt different strategies for answering items, such as solution behavior and rapid guessing behavior. In high-stakes tests, examinees may resort to rapid guessing for difficult and complicated items due to time pressure. In addition, examinees with low motivation may constantly switch between solution behavior and rapid guessing behavior to complete the test. Solution behavior signifies thoughtful consideration, while rapid guessing suggests minimal effort. This paper proposes a mixture hierarchical model that utilizes responses and response times to detect rapid guessing behavior in cognitive diagnosis assessments. The model enables the determination of whether examinees have mastered the required knowledge attributes. A Markov chain Monte Carlo method is introduced for estimating the parameters of the model. Simulation studies show that the parameter estimates obtained using this new model are more accurate than those obtained using a non-mixture model under various conditions. Bayesian assessment criteria are employed to evaluate the model fitting. Finally, a practical application is presented to illustrate the utility of the proposed model.