Authors:
Zermiani, Francesca, francesca.zermiani@uv.es, University of Valencia
Mézière, Diane, diane.meziere@utu.fi, University of Turku
Kaakinen, Johanna, johkaa@utu.fi, University of Turku
Vargas, Cristina, Cristina.Vargas@uv.es, University of Valencia
Salmerón, Ladislao, ladislao.salmeron@uv.es, University of Valencia
Keywords: Eye-movements, reading comprehension, meta-analysis
Abstract:
Reading comprehension is a cognitive process that involves the complex interplay of reader, text, and task characteristics, ultimately leading to the construction of mental models and meaning from a given text. Assessing this process is highly relevant in educational contexts, where reading proficiency is a fundamental skill to acquire and enables students to learn efficiently across disciplines. Eye-movement metrics, obtained from eye-tracking technology, have been widely investigated as indicators of the cognitive processes underlying reading. In particular, prior research has shown that eye-movement metrics, such as fixation times or number of regressions, can be used to predict performance on a variety of comprehension tasks. However, the strength and direction of this relationship across different eye-movement measures remain unclear, and eye-movement indicators of good or poor comprehension have not yet been clearly identified. In this project, we run a meta-analysis to systematically synthesize existing research and determine which eye-movement measures are the most useful predictors of reading comprehension. We will conduct a systematic literature search across major databases (PubMed, PsycINFO, Web of Science, Scopus, and dblp) to identify studies that examined the relationship between eye-movement metrics and reading comprehension measures. We will only include studies involving silent reading in which longer texts are presented alone, and in which readers do not have access to the text when completing the comprehension assessment. Studies will be coded for sample characteristics (e.g., age, language, cognitive development), text characteristics (e.g., length, type, difficulty), assessment characteristics (e.g., comprehension measurements) and other methodology (e.g., reading device, eye-tracking device, sampling rate). A random-effects meta-analysis will be performed to estimate overall effect sizes for different eye-movement metrics. We will explore potential moderators, such as text and task characteristics, and participant demographics, using meta-regression analyses. Additionally, publication bias and heterogeneity will be assessed to ensure the robustness of our findings. We plan to have preliminary results by the time of this workshop. The findings of this meta-analysis have implications for educational assessment, reading interventions, and the development of educational technologies. Furthermore, the results aim to advance our understanding of the predictive relationship between eye-movement and reading comprehension measures.