ClearType improves our accuracy at recognizing words
Most people agree that ClearType makes text look nicer, especially when tuned. But does it help us read more efficiently? This is a difficult question because adult readers can read at a very fast rate – good readers can read 250-300 words per minute, which is a word every 200-240 milliseconds on average. Measuring something that takes less than a quarter of a second is a challenge.
Tachistoscopic Lexical Decision is a technique that has been used extensively to measure word recognition. Tachistoscopes were a mechanical device for presenting materials briefly, but have been replaced by computers. Readers are briefly shown either a regular word in their native language (e.g. distant) or a pseudoword (e.g. tadints), which is not a word in their native language. The reader’s task is to say if they saw a word or pseudoword. The advantage to this methodology it only requires the reader to complete a very simple task of pressing one button for words or another for pseudowords. This allows more accurate timing for word recognition.
The below animated image demonstrates what the reader sees during this task. The word and pseudoword here are appearing on screen for a longer period than in a typical study, but special software is needed to create shorter presentations.
A team led by Professors Lee Gugerty and Rick Tyrrell at Clemson University studied the effects of ClearType on the tachistoscopic lexical decision task. They asked 24 readers to respond to 280 words and 280 pseudowords. Half of the words were presented with ClearType rendering and half with the regular black & white rendering. They found that readers were statistically reliably more accurate at recognizing words when words were rendered with ClearType.
After correcting for guessing, this translates to a 17% accuracy improvement at recognizing words with ClearType than with black & white. This is a large difference in word recognition performance with a task that simulates the real world situation of looking at words for very brief periods of time.
Kevin Larson
Gugerty, L., Tyrrell, R. A., Aten, T. R. & Edmonds, K. A. (2004). The effects of subpixel addressing on users’ performance and preferences during reading-related tasks. ACM Transactions on Applied Perception, 1(2), 81-101.
