Rethinking reading comprehension: digital situation models in primary education
In recent years, there has been a decline in reading skills among young people in the Netherlands. Due to the persistently disappointing results, criticism of traditional text-with-questions assessment is growing. This type of testing is considered too superficial to measure deep reading comprehension, as test takers who read a text thoroughly often score no better than those who only scan for the parts that answer specific questions. Furthermore, it is difficult to determine which aspects of reading comprehension a pupil is struggling with based on the assessment results.
The project ‘Navigation to better reading comprehension’, in collaboration with Cito, aims to create a digital assessment system that both supports better reading comprehension and more effective testing of it. This will be achieved by applying principles from human-computer interaction and opportunities of AI, thereby contributing to the field of learning technologies.
Mental situation models
Reading comprehension is the ability of gaining knowledge in a targeted manner and happens through the creation of a mental ‘situation model'. A situation model is that what a reader forms of the text that goes beyond the literal text, combining what is stated in the text with what can be inferred, based on one's background knowledge. It is classified as the deepest level of text comprehension, according to the construction-integration theory. With traditional assessment, assessing this deep level of reading comprehension is barely supported.
Instead, pragmatic reading is encouraged which is a superficial approach to reading a text with the sole aim of answering a question. This is partly due to the limitations of traditional tests being paper-based combined with the need for large-scale, fast and simple computer grading. These constraints make it difficult for test designers to create rich assessments, since they are largely restricted to multiple choice questions.
To overcome the shortcomings of traditional reading assessment, this project will focus on studying how mental situation models are formed and how they can be translated into digital representations. The goal is to create a new, digital type of assessment that can measure the deepest level of reading comprehension.
Student
Finola Parkinson
Academic supervisor(s)
- Dr. Matthieu Brinkhuis
- Lientje Maas (CITO)