Readability of hyperlinked data with experiments on school textbooks

Abstract

Readability is the ease with which a text can be read and understood. Readability is a central concept while reviewing texts and directly impacts the utility of a particular text. If the information is available in several pieces of text that are connected by links, the readability is affected. This thesis describes our experiments to study the readability of linked pieces of texts. A user study was conducted on school textbooks by manually adding hyperlinks to some of the central concepts in the text. The study was conducted on school children whose second language was English. It was found statistically significant over plain texts. The results are reported in the thesis. The thesis also describes an automated process of text enrichment, that is, automatically adding hyperlinks to text. The comparison with manually tagged text is reported. Text simplification is also useful to enhance readability. The thesis reports the results of applying text simplification on texts in addition to adding hyperlinks. It comments on the combined effect of simplification and enrichment on comprehensibility and hence readability. Finally it discusses a possible way to measure that change in readability due to linking. A formula is proposed for readability of linked documents that gives a reasonable approximation to the readability and can serve as a baseline for future research.