ABSTRACT
This chapter demonstrates how multiple regression techniques can be used to reading times for sentences in passages. It also examines the extent to which sentence reading times can be predicted by a large number of predictor variables, which are properties of words, sentences, and passages. The chapter examines the extent to which readers’ comprehension scores be predicted by the readers’ allocation of processing resources to various units, components, and dimensions of text during reading. It also discusses issues and problems in applying multiple regression in addition to reporting some substantive results that can be obtained. One virtue of methodology is that can apply it to any passage. The comprehension test contained 120 3-alternative, forced choice (3AFC) questions. There were 10 equations for each of the 12 passages. The questions were sufficiently challenging to test for comprehension, as opposed to simple word recognition. The computer presented the questions and recorded the subjects’ answers.
