Large-Scale Study of an Automated Data Collection Method of a Computer Assisted Instruction ABA Program

Background: Computer assisted instruction (CAI) has long been presented as beneficial to learning. Literature supports CAI as it relates to faster learning, positive attitudes towards computer usage and CAI, and increased motivation (Kulik, 1994; Flether-Flinn & Gravatt, 1995). Other benefits include generalization (Bosseler & Massaro, 2003; Hetzroni & Shalem, 2005), fast and efficient data collection, and automatic data reporting.

Previous research has focused on comparing CAI’s data collection methods and other data collection including paper and pencil methods (Couper, 2005; Velikova, et, al, 1999). Additionally different automatic data collection methods have also been compared in regards to their usability, data entry, and reliability. However, the number of studies that analyze the structure and the types of information that automated data collection methods of CAI yield is very limited and needed in this growing field. This study looked at the automated data collection method of TeachTown, a CAI, to understand the underlying structure of the data reporting, important variables, and the types of results that can be collected by using this method.

Objectives: To understand trends, uncover underlying structures, and extract important variables from an automated data collection method from TeachTown, a CAI. TeachTown: Basics is an Applied Behavior Analysis (ABA)-based CAI that was designed to meet the developmental needs of children with Autism Spectrum Disorders in the 2-7 year developmental range.

Methods: Data from over 3000 students between the ages of 1-20 years of age currently using TeachTown was examined using an exploratory data analysis to understand general patterns, important variables, and the automated data collection method. Areas examined were student’s usage of the software, frequency of attempts, progress, and specific domain information (response time, accuracy, and prompting). Additionally a case analysis was conducted on a single student to better understand individual patterns in software usage and learning.

Results: After analyzing over 2 million data points, results point at the underlying structure of the automated data collection method. Important variables related to students’ scores in the software are reviewed, as well as a description of a frequent user are examined.

Conclusions: Results will allow future CAI to make improvements in their products, customer support system, and in their automated data collection system. Limitations and directions for future research are also discussed.