Word per Minute: The Fallacy of the Wilkins Rate of Reading Test

A great article appeared in Guardian today which you can find here. The piece discusses a paper which has appeared in the journal Psychological Science in the Public Interest which you can download here. The good news is that it is freely available and not hidden behind a paywall, the bad news is that if you hoped that you could learn to read faster, while maintaining comprehension and accuracy, through speed reading courses you are in for a disappointment. Woody Allen put it

succinctly when he said 'I took a speed reading course and read War and Peace in 20 minutes ....... er it involves Russia' - I can't do comedic pauses in print so this probably does not do justice to the humour of his remark. Like much humour it carries quite a profound message in this case about some the trials of coloured filters to treat reading difficulty.
Returning to the scientific paper, the authors review the evidence for training to increase reading speed through speed reading apps, and technology introduced for smart phones and other digital devices. What they show is that you can not double or triple reading speed whilst conserving other aspects of reading including accuracy and comprehension. In short, there is trade off between reading speed and other aspects of reading.
They go on to say that reading speed is more related to language processing skills so that if information comes in faster than the 'comprehension system' can process it - the increased speed is wasted. Practice with language is what is required to become a better reader. We all know this. When we have to read a document in a sphere we are familiar with we can read more quickly because we spend less time over familiar words. When we have read outside our subject area reading slows - sometimes dramatically.
So what does this have to do with trials of coloured lenses and overlays to improve reading?  A common diagnostic test and end point for these trials is the Wilkins Rate of Reading test (see blog post for September 2015). The WRRT consists of randomly ordered high-frequency words with no syntax or meaning. This means that only speed and to a limited extent accuracy can be measured. As result, even if we ignore the other sources of bias in these trials we can not be sure that this improvement in reading results from visual perceptual factors or simply a trade off between the rate of reading and comprehension. Even if we accept that the WRRT is a useful diagnostic test for so called visual stress it is certainly not an appropriate dependent variable for clinical trials of coloured lenses and overlays.