The research unit of Roma Tre University – Department of Philosophy, Communication and Performing Arts – has identified as the object of its investigation the strategies through which readers integrate text with additional content—web searches, multimedia materials, encyclopedic insights, iconographic resources—using them to strengthen comprehension, broaden the context, or enrich the interpretation of the text, recognizing how this phenomenon profoundly modifies the reading experience, transforming it into a hybrid analog-digital activity that is highly interactive and often collaborative.
The study utilized a mixed-methods experimental design, combining a paper-based reading and underlining test with an on-screen reading test of the same texts monitored via Eye-Tracker. The objective was to identify textual triggers that induce readers to interrupt their reading to search for, deepen, or contextualize information, verifying how such behaviors are influenced both by the characteristics of the text (genre, density of implicit intermedial references) and those of the reader (age, experience, education).
The sample included three reader profiles: upper secondary school students (15–18 years old), university students (18–35 years old), and adult readers (over 35 years old). For the test, the initial passages (freely available online or with the rights holder’s authorization) of four texts were selected: two narrative and two non-fiction, two with a low density of external references (A Winter’s Promise. The Mirror Visitor 1. by Christelle Dabos and A Decent Life. A Morality for All of Us by Todd May) and two with a high density of references (The Years by Annie Ernaux and La fidanzata automatica by Maurizio Ferraris).
In the paper-based test, participants underlined trigger words and passages for different types of external searches and reported any pauses using specific methods provided to them before starting the reading, making the conscious dimension of augmented reading visible. The collected data were examined through ATLAS.ti, a computer-assisted qualitative analysis software designed to manage, organize, and interpret complex textual, visual, or multimedia data. Through advanced coding, classification, and visualization tools, ATLAS.ti allowed us to identify recurring patterns and compare the reading behaviors of the different groups.
The on-screen reading test of the same texts with an Eye-Tracker was administered within the CoSMIC Lab (Cognition, Social Multimodal Interaction and Communication) of Roma Tre University, in collaboration with the researchers working at that facility.
In particular, Eye-Tracking made it possible to observe unconscious reading mechanisms by measuring eye movements. Thanks to an infrared camera for pupillometry, the Eye-Tracker allowed for the recording of fixations, saccades, and regressions during the act of reading by the participating subjects. Specifically for the NEREIDE project, an experimental design was formulated that exposed each subject to the four different types of texts. Each participant—after signing the informed consent and receiving instructions from the researchers—took part in the experimental session in the laboratory: following a brief camera calibration phase, they were invited to read the four texts, projected on a monitor, silently at their own pace. The texts were administered in a randomized form to avoid order bias. Furthermore, to avoid excessive eye strain, a gray background and a sufficiently large font were chosen to ensure accurate measurement. For the initial analyses, a qualitative investigation was conducted through heatmaps, which highlighted several exploratory trends consistent with the starting hypothesis.
The comparison between the two levels of analysis showed a strong convergence: the words identified as triggers in the paper-based test very often coincide with the areas of highest attentional density in the heatmaps. Texts characterized by a higher density of intertextual references show both more underlining and a more intense attentional distribution, suggesting a higher cognitive load.
Overall, the frequent overlap between conscious markings and implicit attentional patterns confirms the effectiveness of the integrated approach adopted.
All data collected during the two tests, anonymized, are available in open form within the dataset section of the site and are listed below.
Datasets
Comparative Analysis
The comparative analysis relates the results of the reading and highlighting task on paper with those of on-screen reading monitored through eye-tracking. The aim of the analysis is to compare readers’ conscious behaviors, expressed through their highlights, with the attentional data recorded by the eye-tracker. The analysis includes 109 valid trials for both experiments out of 132 administered, selected according to the quality criteria required by eye-tracking technology.
ANALYSIS OF THE READING AND HIGHLIGHTING TASK ON PAPER
This analysis focuses exclusively on the data collected from the reading and highlighting task on paper. The study includes 129 valid trials out of 132 administered and directly observes readers’ augmented reading behaviors, using the highlights made on the texts as indicators. The results take into account both the participants’ sociodemographic profile and the order in which the two tasks included in the experimental design were carried out.
