The Increasingly Inextricable Relationship between Orthographic and Phonological Coding in Learning to Read: Some Reservations about Current Methods of Operationalizing Orthographic Coding

Unlike phonological coding, orthographic coding as a psychological construct is not well understood and has not been well defined. Currently, the most often used tasks to measure orthographic coding include orthographic choice (“which is a real word?”—rain, rune), homophone choice (“which is a number?”—ate, eight), spelling recognition (time, tient, tihm, teirn) and, more recently, spelling from dictation. In the present chapter, we express our reservations about using these measures because, in our opinion, they evaluate word identification and/or spelling ability rather than orthographic coding as a basic cognitive process that underlies word identification and spelling ability. We present research findings from our laboratory and review research done elsewhere to support our reservations. We also point out the risk of using only reading level match designs in evaluating orthographic coding effects. Using our data, we show that matching samples based on one reading variable creates a mismatch on another, and that interpretations of the findings based on exclusive employment of this type of design are inevitably problematic. We also discuss research that shows that matching poor and normal readers on tests that depend heavily on word specific knowledge, in particular tests of word identification and/or spelling ability, will tend to equate these groups on orthographic coding tasks that also depend heavily on word specific knowledge, while matching them on tests that depend heavily on pseudoword decoding ability will tend to equate them on orthographic coding tasks that also depend heavily on general orthographic knowledge.