Read, Understand, Learn, & Excel: Development and Testing of an Automated Reading Strategy Detection Algorithm for...

Area of Science:

• Cognitive Psychology
• Educational Technology
• Human-Computer Interaction

Background:

• Reading comprehension is crucial for academic success.
• Students use various strategies to process text for learning.
• No existing tools automatically detect these reading strategies.

Purpose of the Study:

• Develop an automated tool, Read, Understand, Learn, & Excel (RULE), to detect reading strategy use.
• Assess RULE's accuracy in identifying strategies during digital, expository text reading.

Main Methods:

• Iterative design to create a computer algorithm for 9 reading strategies.
• 12 undergraduates read digital expository texts.
• Comparison of computer-detected strategies against human observer documentation.

Main Results:

• The RULE tool achieved 75% or higher agreement with human observers for 6 out of 9 strategies.
• Lower agreement (<60%) was observed for previewing, evaluating remaining text, and summarizing strategies.
• The study demonstrated proof of concept for automated reading strategy detection.

Conclusions:

• The RULE tool offers an objective method for capturing readers' engagement with academic text.
• This technology has potential clinical implications for understanding reading behaviors.
• Further refinement is suggested to improve the tool's sensitivity.