Research is constantly engaged in finding more productive and powerful ways to support quality learning and teaching. However, although researchers and data scientists try to analyse educational data most transparently and responsibly, the risk of training machine learning algorithms on biased datasets is always around the corner and may lead to misinterpretations of student behaviour. This may happen in case of partial understanding of how learning log data is generated. Moreover, the pursuit of an ever friendlier user experience moves more and more Learning Management Systems functionality from the server to the client, but it tends to reduce significant logs as a side effect. This paper tries to focus on these issues showing some examples of learning log data extracted from Moodle and some possible misinterpretations that they hide with the aim to open the debate on data understanding and data knowledge loss.

翻译：虽然研究人员和数据科学家试图以最透明和负责的方式分析教育数据,但是,在偏向数据集方面,培训机器学习算法的风险总是近在眼前,并可能导致对学生行为的误解。这可能发生在对如何生成学习日志数据有部分理解的情况下。此外,对日益友好的用户经验的追寻会让用户越来越多地从服务器到客户,但往往会减少重要的日志作为副作用。本文试图重点阐述从Moodle中提取的学习日志数据的一些实例,以及他们为了开启关于数据理解和数据知识损失的辩论而隐藏的一些可能的误解。