Assessing text quality as an indication of underlying skills still remains challenging; irrespective of the approach, many studies struggle with reliability or validity problems. If writing is considered problem-solving, a report must make the reader understand the described situation and call for its mental reconstruction. Therefore, text quality may not only comprise linguistic aspects but also the cognitive-functional power of a text. The presented study aims at exploring the functionality of students’ reporting texts in relation to general text-quality measures, using a corpus of accident reports written by German fifth- and ninth-graders (n = 277) prompted by a pictorial stimulus of a bike accident scenario. An online tool was developed in which 277 university students graphically reenacted the situation from one respective text according to the existence, position, and color of the involved elements. Thereafter, the match of the resulting spatial reconstructions with the original situation was assessed by two raters. While most subscales showed sufficiently high interrater reliabilities, the aggregated functionality score (α = .74) had medium-high correlations with other text-quality ratings and was comparably dependent on grade, education level, and linguistic family background. However, the correlational pattern, regression analysis, and factor analysis showed that the functionality score also contributed unique portions of variance to the assessment of writing skill that were not represented by rating measures. Moreover, the direct indication of whether a text allows for the reader’s adequate cognitive representation is evident. Altogether, the approach of indicating text functionality through practical understanding offers a sound, though empirically laborious, alternative for text-quality measurement. Results are discussed with regard to the didactical strategy according to which students can improve their writing when they observe whether others can make use of their texts.

中文翻译：

---

文本质量的直接功能衡量：读者理解吗？

评估文本质量作为潜在技能的指标仍然具有挑战性；无论采用何种方法，许多研究都面临着可靠性或有效性问题。如果写作被认为是解决问题，那么报告必须让读者理解所描述的情况并呼吁其进行心理重建。因此，文本质量可能不仅包括语言方面，还包括文本的认知功能能力。本研究旨在利用德国五年级和九年级学生 (n = 277) 在自行车图片刺激下撰写的事故报告语料库，探索学生报告文本与一般文本质量衡量标准相关的功能。事故场景。开发了一种在线工具，277 名大学生根据相关元素的存在、位置和颜色，以图形方式重现了各自文本中的情况。此后，由两名评估者评估最终空间重建与原始情况的匹配度。虽然大多数子量表显示出足够高的参与者间可靠性，但聚合功能得分 (α = .74) 与其他文本质量评级具有中高相关性，并且相当依赖于年级、教育水平和语言家庭背景。然而，相关模式、回归分析和因子分析表明，功能得分还对写作技能评估产生了独特的方差部分，而这些方差并未通过评分指标来体现。此外，直接表明文本是否允许读者进行充分的认知表征是显而易见的。总而言之，通过实际理解来指示文本功能的方法为文本质量测量提供了一种合理的（尽管在经验上费力的）替代方案。讨论了关于教学策略的结果，根据该策略，当学生观察其他人是否可以使用他们的课文时，他们可以提高他们的写作水平。