The answer, according to the findings, is a clear no. Among the fifth-grade children who participated, stronger comprehension went hand in hand with markers of reading fluency. Children who understood the texts better tended to make fewer fixations overall, produced fewer regressions back to previously read material, and executed longer saccades, meaning their eyes jumped farther forward with each movement. These are classic signatures of an efficient, automatic decoding process. In contrast, the adults in the sample, who included twenty undergraduates and nine faculty members, showed a different pattern. For them, better comprehension was associated with slower reading, more fixations, and generally more attentive, effortful engagement with the text. Rather than fluency signaling understanding, in adults it was deeper cognitive investment that predicted how well the material was grasped.

This developmental reversal makes sense when situated within established theories of reading acquisition. The influential framework proposed by LaBerge and Samuels in the 1970s holds that reading depends on automaticity: when lower-level processes such as letter recognition and word decoding become automatic, attentional resources are freed for higher-level comprehension. For children still consolidating these lower-level skills, the efficiency of basic word recognition remains a bottleneck. A child who must laboriously assemble words letter by letter has little capacity left to build a coherent mental model of the passage. Consequently, in young readers, fluency measures function as a reasonable proxy for comprehension potential, a view consistent with the large body of evidence showing that text reading fluency mediates the relationship between decoding skill and comprehension throughout the elementary school years.

By adulthood, however, decoding is typically fully automatized for skilled readers, and variation in fluency among competent adult readers no longer constrains understanding in the same way. Instead, comprehension differences among adults reflect what happens after individual words are recognized: integrating sentences into a discourse model, drawing inferences, monitoring understanding, and relating new information to prior knowledge. Walter Kintsch’s construction-integration framework describes comprehension as the construction of a textbase and a situation model, processes that demand deliberate cognitive engagement. The finding that adults who read more slowly and fixated more often comprehended better suggests that these adults were allocating extra time to exactly these integrative operations. Slower reading in a skilled adult is not a sign of weakness but often a sign of deeper processing, particularly when texts are expository and informationally dense.

The distinction maps onto the difference between lexical and post-lexical processing. Fixation durations and word-based measures during a first pass through a sentence largely reflect lexical access and syntactic parsing, whereas regressions, second-pass reading, and selective reinspection often reflect repair, integration, and inference generation. In children, the study’s correlational pattern suggests that the first-pass efficiency measures dominate the comprehension picture. In adults, the engagement measures, including the willingness to slow down and dwell on the text, take over as the meaningful correlates of understanding. This shift from reliance on low-level skill to reliance on high-level processing represents a developmental progression in the very architecture of comprehension.

Several features of the study’s design merit attention when interpreting these results. Participants read age-appropriate expository texts, which means children and adults were not reading identical material. This choice, while necessary to avoid floor and ceiling effects, means the developmental comparison concerns comprehension of suitably calibrated texts rather than comprehension of a single fixed passage. Expository texts were a sensible choice because they place heavy demands on background knowledge and inference, and they resemble the kinds of reading required in educational settings from middle childhood onward. Eye-tracking data were collected during the participants’ first read of each text, which is important because rereading changes eye-movement patterns substantially; prior work by Hyönä and Kaakinen has shown that fixation durations drop and patterns reorganize when a text is encountered a second time. Capturing the first encounter ensures that the recorded eye movements reflect genuine online comprehension processes rather than memory-supported review.

The sample itself was modest but appropriate for a correlational study of this kind: twenty-one fifth-graders and twenty-nine adults. All participants were native Spanish speakers, which matters for several reasons. Spanish has a relatively shallow orthography, meaning that grapheme-to-phoneme correspondences are largely regular and decoding is typically mastered earlier and more completely than in deep orthographies such as English. Meta-analytic work, including the recent analysis by Leachman, Wolters, and Kim, has demonstrated that the relationship between text reading fluency and comprehension varies systematically with orthographic depth and developmental phase. In a transparent orthography, the fluency-comprehension link in children may reflect a somewhat different balance of skills than it would in English, where decoding struggles persist longer. The fact that fluency still emerged as the key correlate of comprehension in Spanish-speaking children underscores how central automatized word processing remains for comprehension even when the writing system is relatively friendly to the learner.

The study also connects to a broader movement in reading science away from overly simple models. The classic Simple View of Reading proposed by Gough and Tunmer in 1986 framed comprehension as the product of decoding and linguistic comprehension, a formulation that has been enormously productive but increasingly recognized as incomplete. More recent integrative models, such as Kim’s Direct and Indirect Effects Model of Reading and the framework advanced by Duke and Cartwright, emphasize that reading comprehension rests on a constellation of skills, including executive functions, vocabulary, background knowledge, and inference-making, whose relative contributions shift across development. The eye-tracking evidence from this study provides a behavioral, moment-to-moment signature of exactly such a shift: the same observable behavior, such as a high fixation count, carries different implications about comprehension depending on the reader’s developmental stage.

Executive functions deserve particular mention in this context. Butterfuss and Kendeou have argued that working memory, inhibition, and cognitive flexibility support the construction and maintenance of a coherent situation model, especially during the integration of information across sentences and paragraphs. The adult pattern observed here, in which more fixations and slower reading accompanied better comprehension, is consistent with readers strategically deploying these executive resources. Similarly, research on prior knowledge by Kaakinen and colleagues has shown that readers allocate longer fixations to text segments that are relevant to their reading goals, particularly when background knowledge must be actively coordinated with incoming information. Adults who engage in this kind of selective, effortful reading are, in effect, trading speed for depth, and the comprehension measures reward that trade.

Practically, the findings carry implications for assessment and instruction. For children in the intermediate grades, the results reinforce the value of building reading fluency, not as an end in itself but as the gateway to comprehension. Repeated reading, wide independent reading of connected text, and instruction that promotes automatic word recognition all serve to free cognitive resources for meaning-making. For adolescents and adults, especially those in teacher education programs where concerns about reading comprehension demands have recently been raised in the literature, the results suggest that interventions should target higher-level skills: inference generation, discourse integration, metacognitive monitoring, and the strategic slowing down that deep texts require. A one-size-fits-all approach that treats slow, effortful reading as a problem to be corrected would, for adult readers, mistake engagement for inefficiency.

The study’s methodological choices also illustrate good practice in eye-tracking research. The authors addressed outlier detection and treatment following established statistical guidance, an important step given that fixation distributions are typically skewed and sensitive to occasional anomalous measurements. The decision to include both undergraduate students and faculty professors within the adult group introduced a range of reading expertise among the adults, which may have sharpened the contrast between fluency-driven and engagement-driven comprehension profiles. The openness of the data on the Open Science Framework allows other researchers to verify the analyses, test alternative models, and pool evidence across studies, practices that strengthen the cumulative reliability of the field.

Some caveats remain. The correlational design cannot establish causal direction: it is possible that children who comprehend better become more fluent, rather than fluency driving comprehension, although longitudinal and intervention evidence elsewhere supports the latter pathway as well. The modest sample sizes, while adequate for regression-based analyses according to minimum sample size guidance, limit the precision of estimates and the ability to detect interactions. The inclusion of professors alongside undergraduates raises questions about whether the adult pattern is homogeneous or driven by the most expert readers. And because comprehension was measured after a single reading, the study cannot speak to how readers strategically reread or study text over time.

Nevertheless, the central contribution stands out clearly: the eye movements that betray comprehension are not universal but developmental. What looks like skill in a child, namely swift, forward-moving, uninterrupted reading, looks different in an adult, for whom comprehension is signaled by deliberate, resource-intensive engagement with the text. Understanding this progression helps researchers build more accurate models of reading development, helps educators match instruction to the reader’s current stage, and reminds us that the path from fluent decoding to deep understanding is not a single leap but a gradual reorganization of the entire reading system.

Subject of Research: Comprehension is in the Eye of the Reader: An Eye Tracking Study of Children and Adults

Article Title: Comprehension is in the Eye of the Reader: An Eye Tracking Study of Children and Adults

Article References: Tabullo, Á., Jofre, E. M. G., Del-Punta, J. A., Rodríguez, K. V., & Gasaneo, G. (2026). Comprehension is in the Eye of the Reader: An Eye Tracking Study of Children and Adults. Trends in Psychology. https://doi.org/10.1007/s43076-026-00537-4

Image Credits: AI Generated

DOI: 10.1007/s43076-026-00537-4

Keywords: Comprehension, Reader, Tracking, Children, Adults, scientific research