Rapid online assessment of reading ability

Area of Science:

• Cognitive psychology
• Educational psychology
• Psychometrics

Background:

• Accurate reading ability models require large, diverse samples.
• Standardized reading tests are resource-intensive and difficult to administer.
• Web-based assessments can overcome these limitations.

Purpose of the Study:

• To evaluate a novel, self-administered lexical decision task (LDT) as a measure of reading ability.
• To determine the reliability and validity of this online assessment.
• To develop an efficient tool for large-scale reading research.

Main Methods:

• A two-alternative forced choice, time-limited lexical decision task (LDT) was developed for web-browser delivery.
• Performance on the LDT was compared with scores from standardized reading tests (e.g., Woodcock-Johnson).
• Item response theory was used to optimize word and pseudoword lists for a short experiment (76 trials).

Main Results:

• LDT performance strongly correlated with standardized reading ability measures (r=0.91, disattenuated r=0.94).
• The LDT demonstrated high reliability (r=0.97).
• An optimized 2-3 minute LDT experiment yielded a reliable reading ability measure (r=0.95).

Conclusions:

• The self-administered Rapid Online Assessment of Reading (ROAR) is a valid and reliable measure of reading ability.
• ROAR overcomes the limitations of traditional, in-person assessments.
• This tool enables efficient, automated, large-scale online research into reading mechanisms.