Three Related Experimental Pilot Studies Aimed to Innovate Reading Intervention for Students with Word Reading Difficulties

Abstract This paper summarizes three theoretically related experimental pilot studies conducted in developing a new intervention for students with word reading difficulties. Each independent experiment tested instructional elements influenced by perspectives on lexical quality, connectionism, and statistical learning. Intervention components that showed promise moved on to become part of a base program tested in a subsequent study. In each study, students in grades 2–4 with low word reading skills were randomly assigned to one of two intervention conditions. Researcher‐supported tutors provided small‐group instruction five days per week over ten weeks at under 25 min per session. Treatment fidelity exceeded 90% in each experimental study. Standardized and researcher‐developed measures of word and text reading were administered pretest and posttest. Although few statistical differences were observed, several educationally meaningful effect sizes were observed between groups across each study. In Study 1, effect sizes favored a condition in which decoding instruction and practice included variability in spelling–sound correspondence and larger letter units, compared to a condition in which standard spelling–sound pronunciations and sound‐by‐sound decoding were emphasized. In Study 2, effect sizes favored a condition in which a spelling activity was embedded in decoding instruction compared to identical decoding instruction without spelling. In Study 3, contrary to our expectations, embedding semantic information within a decoding and spelling program was less effective than the same instruction without semantic information. We discuss the results related to innovations to improve generalized word reading skills and our next steps that involve a randomized controlled trial and further intervention development.