Assessing the Lexico-Grammatical Characteristics of a Corpus of College-Level Statistics Textbooks: Implications for Instruction and Practice

Abstract

A corpus of current editions of statistics textbooks was assessed to compare aspects and levels of readability for the topics of measures of center, line of fit, regression analysis, and regression inference. Analysis with lexical software of these text selections revealed that the large corpus can be described well by three index variables that summarize the lexical and grammatical complexity of the textbook excerpts. Assessment of those three variables indicates that substantial differences exist in the readability of the topics within textbooks with respect to lexical and grammatical complexity. This analysis suggests that general readability of introductory statistics topics within textbooks varies substantially and it is a recommendation that instructors: (1) be prepared to provide additional support for topics that are more grammatically and lexically complex, and (2) be aware that they can input their instructional materials into LexTutor VP or Coh-metrix as a quick screen for possible readability issues.