计算心理测量视域下的学生数字素养测评

Evaluation of Students’Digital Literacy from a Computational Psychometric Perspective

定期开展数字素养发展监测与评估是提升全民数字素养的基础和前提。当前,学生数字素养评价主要沿用“依据规则给潜在特质赋值”的经典测量观,以标准化测试及自陈式量表为主要测评工具,存在评价内容有限、评价方法片面等问题。计算心理测量理论为破解学生数字素养评价难题提供了新的理论和方法支持。它主张综合运用心理测量学、计算机科学等多学科的研究方法,将评价嵌入真实情境任务中,激发并追踪主体与情境之间的交互,通过多源异构数据采集与分析,实现过程性评价。计算心理测量视域下的学生数字素养测评框架基于自上而下理论驱动和自下而上技术驱动相结合的推理链,在以证据为中心的设计思想指导下,构建复杂的任务情境获取细粒度数据,并利用数据挖掘技术从海量细粒度数据中提取证据,进而实现精准评价。以此框架开展的面向小学高年级学段的数字素养测评实践表明,该测评范式有助于形成从数字素养“能力模型”到“测量模型”的闭环推理链,在经过多轮次迭代优化后,可形成高质量的学生数字素养测评任务、可靠的学生数字素养证据和稳定的学生数字素养测量模型。

Regular monitoring and assessment of digital literacy development is the foundation and premise for enhancing the digital literacy of the entire population.Currently,the evaluation of students’digital literacy primarily adheres to the classical measurement view of“assigning values to potential traits based on rules”,mainly utilizing standardized tests and self-reported scales as assessment tools.This approach faces limitations such as restricted evaluation content and one-sided evaluation methods.Computational psychometrics offers new theoretical and methodological support to solve the challenges in evaluating students’digital literacy.It advocates for the combined use of research methods from multiple disciplines,including psychometrics and computer science,embedding assessment in real-world tasks to stimulate and track interaction between subjects and contexts.Through the collection and analysis of multi-source heterogeneous data,a process-oriented evaluation is realized.The evaluation framework for students’digital literacy from the computational psychometrics perspective,guided by an evidence-centered design philosophy,constructs a reasoning chain combining top-down theoretical drive and bottom-up technology drive.It builds complex task scenarios to gather fine-grained data and uses data mining techniques to extract evidence from vast amounts of detailed data,thereby achieving precise evaluation.The practice of digital literacy assessment for upper elementary grades,conducted with this framework,demonstrates that this assessment paradigm helps form a closed-loop reasoning chain from the“capability model”of digital literacy to the“measurement model”.After multiple iterations of optimization,it can produce high-quality digital literacy assessment tasks,reliable evidence of students’digital literacy and a stable measurement model.