Lexiland: A Tablet-based Universal Screener for Reading Difficulties in the School Context

Abstract

Massive and timely screening of the student population for early signs of reading difficulties is needed to implement timely effective remediation of these difficulties. However, traditional approaches are costly and hard to apply. Here, we present Lexiland, a tablet-based reading assessment tool for kindergarten and primary school children developed to be applied in school settings with minimal personnel intervention. Following a story line, players help a character of the game perform several tasks that measure different predictors of reading outcomes. Most of the tasks that usually involve a verbal response were switched to receptive tasks to demand a touch-screen response only. The tablet application was administered to a sample of N = 616 5-yo kindergarten children and to a sub-sample of these children twice during the following two years (First and Second Grades). Applying logistic regression and cross-validation, we selected a reduced subset of tasks that can predict with great sensitivity and specificity, whether a five-year-old child will have reading difficulties by the end of first grade (sensitivity 90% and specificity 76%) and two years later (sensitivity 90% and specificity 61%). Importantly, Lexiland is a scalable tool to implement universal screening, given the increasing availability of devices able to run android and iOS applications.