Rapid Online Assessment of Reading

The Rapid Online Assessment of Reading ability (ROAR) is a completely automated assessment of automaticity in word recognition that runs in the web-browser. Through a series of validation studies, our research has validated that scores on the ROAR are consistent with scores on other standardized reading assessments, such as the Woodcock Johnson Basic Reading Skills index, that involve reading a sequence of words out loud. Moreover, scores on the ROAR provide a highly reliable index of reading ability in participants as young as first grade. Read more about the development and validation of the ROAR in Nature Scientific Reports:

Yeatman, Jason D., Kenny An Tang, Patrick M. Donnelly, Maya Yablonski, Mahalakshmi Ramamurthy, Iliana I. Karipidis, Sendy Caffarra, et al. 2021. “Rapid Online Assessment of Reading Ability.” Scientific Reports 11 (1): 6396.

Details of the ROAR

The ROAR is a Javascript web application that will run on any personal computer through the web-browser. It measures word-recognition abilities by rapidly presenting a sequence of real words and made up words (“pseudowords”) and asking the participant to indicate whether each word is real or fake with a keyboard response. The stimuli are calibrated to span a wide difficulty range and, thus, provide an accurate index of word recognition abilities in 1st through 12th grade. Since each trial is a quick decision, the ROAR is able to accurately measure word recognition abilities relatively quickly (5-10 minutes). The ROAR is embedded in a game-like animated environment. Instructions are narrated by a character in the game and there are practice trials with feedback to ensure that each participant understands the task. In our validation studies, we found that children as young as second grade were able to complete the ROAR without a proctor or test administrator and children in first grade could complete the ROAR with minimal assistance.

Interpreting a ROAR score report

The ROAR is a measure of word recognition. Word recognition is at the foundation of reading and is a prerequisite for reading fluency and comprehension. The ROAR Raw Score is the number of real/fake words that the participant accurately identified. There are 252 items in total (84 per test form) and since each trial has two alternatives (real/fake) random guessing will result in roughly 50% correct. Thus, ROAR Raw Scores range from 126 (chance) to 252 (perfect). High scores mean that the participant is able to rapidly recognize difficult and complex words.

ROAR Standard Scores are created by converting the raw scores to a Woodcock Johnson equivalent score since this is a widely used assessment. These standard scores are normalized by age meaning that each participant’s score is now comparable to others of their same age. Standard scores have a population mean of 100 and standard deviation of 15 based on a normative national sample.

ROAR Percentile Scores give each participants percentile rank relative to a normative national sample. Thus percentile scores can be used to interpret where a students word recognition skills are relative to others of their age.

ROAR Risk Scores classify each participant into one of three three categories: At Risk, Some Risk, or Doing Well. These categories are created by applying cut scores to the percentile rank scores. More specifically, participants performing below the 25th percentile are flagged as At Risk. Students performing between the 25th and 50th percentile are flagged as being at Some Risk. Finally, participants performing above the 50th percentile are classified as Doing Well. The participants within the first two categories performed below average, according to their percentile scores, and may benefit from additional reading support and intervention.

It is important to note that we are currently researching how to best refine the thresholds for identifying students who may benefit from additional support. Thus, it is important that these classifications are interpreted in the context of existing knowledge of each participant’s reading ability.