Abstract
Increasingly, the field of reading instruction is calling for “authentic”, “meaningful”, or “naturalistic” assessment practices. At the same time, considerable research supports explicit instruction and assessment of decoding skills in reading, especially for low-achieving students. There is obvious need for a naturalistic test of decoding skills that is also valid and reliable, yet no such test yet exists. This study examined some technical qualities of a Diagnostic Names Task which promises to help fill this need. An instrumentation study was conducted with a group of 25 elementary grade disabled readers to answer questions about subtest reliability, sensitivity to change due to instruction, meaningful clustering of subskills, and criterion-related validity. Results were generally encouraging, despite a few sampling and design problems. A large-sample study is recommended next to help further improve subtest reliability.
Get full access to this article
View all access options for this article.
