Estimating readability with the Strathclyde readability measure

Despite their significant limitations, readability measures that are easy to apply have definite appeal. With this in mind, we have been exploring the prospects for more insightful measures that are computer-based and, thereby, still easily applied. The orthodox reliance on intrinsic syntactic features is an inherent limitation of most readability measures, since they have no reference to the likelihood that readers will be acquainted with the constituent words and phrases. To accommodate this feature of 'human familiarity', we have devised a metric that combines traditional factors, such as Average Sentence Length, with a measure of word 'commonality' based upon word frequency. This paper details the derivation, nature and application of the Strathclyde Readability Measure (SRM).