Beyond Total Time Measures of Rapid Automatized Naming (RAN): Do Pause Time Variability and Alternate Task Lengths Improve Prediction of Future Reading?

Purpose: Although rapid automatized naming (RAN) is typically measured by the total time to name an array of items, more precise methods may be needed to capture the full range of processes underlying RAN—including pause time, a component of total time that reflects response preparation before the articulation of each item. This study investigated whether two underexplored aspects of RAN—pause time variability (PTV) and task length—can improve or streamline the prediction of future reading difficulties. Using mixed-effects location scale models (MELSMs), which capture item-by-item variability rather than aggregating data, we were able to uncover nuanced patterns in item-level RAN responses.Method: We utilized a longitudinal sample of 142 English-speaking kindergarteners (47.9% girls) oversampled for dyslexia risk. Item-by-item variability and change in RAN pause times were modeled using MELSMs. We then compared model fit between models with PTV and/or total time from kindergarten RAN as the predictor(s) of second grade reading rate, accuracy, and comprehension. Results: Although higher PTV significantly predicted subsequent lower reading rate, accuracy, and comprehension, PTV did not significantly improve prediction beyond total time. RAN tasks with fewer total items were worse at predicting all three reading measures compared to the longer (~40-50 item) formats.Conclusions: The current practice of measuring total time instead of PTV is a robust and more parsimonious way to predict reading rate, accuracy, and comprehension in kindergarteners. RAN tasks must be sufficiently long to challenge the limits of automaticity, given the increasing cognitive demands throughout the task.