TEXT COMPREHENSION ACROSS PRINT AND AUDIO: A PERSON-CENTERED MIXED METHODS STUDY

Abstract

The use of the audio medium (e.g., audiobooks and podcasts) is proliferating in everyday and educational contexts. Yet, research investigating text processing in audio compared to the more commonly used print medium is limited in scope. Specifically, the research so far has majorly focused on younger learners or English language learners, narrative genre texts, operationalized comprehension as a unidimensional construct, and used variable-centered analytical techniques. The current mixed methods study aimed to explore text processing across print and audio by focusing on four interrelated dimensions—learner, text, task, and test. I used finite mixture modeling for the quantitative part of the study to identify meaningful reader and listener profiles. Following the identification and validation of profiles, students from the profiles were interviewed to complement and enhance the understanding of the groups. Specifically, the study aimed to investigate differences across the two mediums vis-à-vis learner characteristics, text and test processing behaviors, and comprehension outcomes. A further goal of the study was to identify meaningful and distinct reader and listener profiles by accounting for affective and behavioral variables, and validating the profiles on cognitive variables. Finally, the study aimed to build qualitatively rich descriptions of the quantitatively unearthed profile groups. To address these aims, undergraduate students (n =130) were recruited from human development courses. They completed measures related to self-efficacy and reported their reading and listening habits. Each participant’s screen was recorded as they processed text in print and audio. Text processing behaviors (e.g., scrolling, increasing playback speed) and off-task behaviors (e.g., eating, fidgeting) were coded. Learner-related, text processing, and task variables were used to find meaningful reader and listener profiles. The profiles were validated using prior topic knowledge and comprehension as covariate and outcome, respectively. Students belonging to each profile were invited for interviews (n = 10). The format was a cued retrospective interview, wherein video clips were used to prompt participants. The interviews were transcribed, segmented into utterances, and coded for learner-related, text-related, task-related, and test-related content. Results from the variable-centered analysis revealed that reading print or listening to audio led to similar performance levels on items targeting recall and inference. However, reading print was associated with higher scores on the item assessing the main idea than listening to audio. Results from the mixture modeling and interviews revealed three reader profiles—Distracted Surfers, Labored Harvesters, and Fluent Surveyors—and three listener profiles—Inconsistently Attentive, Inattentive, and Persistently Attentive. The profiles were found to differ qualitatively on strategies, text processing depth, and attention regulation. This study’s contribution is in expanding the research on comprehension across different mediums both in terms of scope and methodologically. The current investigation demonstrates that learner characteristics and text processing behaviors need to be accounted for when studying comprehension with different mediums. Practically, it has implications for practitioners looking to incorporate audio for content delivery in their courses and for instructional designers developing educational technology tools to optimize learning.