Does Systematic Reading Instruction Impede Prediction of Reading a Shallow Orthography?

Article excerpt

Abstract

Different methods of reading instructions have been the subject of controversy. This study examined the influence of systematic phonics vs. non-systematic phonics methods of instruction on the prediction of reading. 443 kindergarten children were tested on phonological awareness, naming speed and visual word matching using the Bielefelder Screening (BISC). Children were retested in grades one and four. Results showed that although the prognostic validity of kindergarten measures was generally low, it was considerably higher for grade one children in classes with non-systematic phonics instruction. Children who received systematic phonics instruction scored significantly higher on measures of phonological decoding as compared to their peers who received less systematic instruction. Implications for the prediction of reading and early screenings are discussed.

Key words: reading instruction; prediction of reading; phonics instruction; phonological awareness; rapid automatized naming

Introduction

Early identification of children at risk for reading difficulties is of particular importance. It allows timely intervention to avoid serious reading problems which could otherwise cause failure during early schooling and later academic career (e.g. Schabmann & Kabicher, 2007; Torgesen & Burgess, 1998).

Previous research in English speaking countries has demonstrated that reading achievement can be predicted by early measures of phonological awareness (PA) and naming speed. PA implies the ability to manipulate large (e.g. syllables) and small (e.g. phonemes) units of spoken language, which helps children to understand the grapheme-phoneme coding principle in written language (e.g. Bradley & Bryant, 1985; Bryant, Bradley, MacLean & Crossland, 1989). Early difficulties in PA hence interfere with the acquisition of alphabetic coding, particularly at the beginning of learning how to read. PA tasks include rhyming, syllable counting, phoneme deletion, phoneme blending, and phoneme detection (for further tasks see Landed & Wimmer, 2000; 2008).

To measure naming speed subjects are asked to recall and name letters, digits, objects, or colours as quickly as possible (also called rapid automatized naming, RAN). So far, no consensus about the nature of relationship between RAN and reading has been found. Some authors argue RAN should be considered part of a more general "phonological processing" construct that reflects the access to phonological information in the long term memory (e.g. Torgesen, Wagner & Rashotte, 1998). However, previous studies showed that RAN accounts for additional variance in reading even when controlling for phonological awareness (Bishop, 2003; Blachman, 1984; Bowers, 1995; Bowers & Swanson, 1991; Compton, DeFries & Olson, 2001; Cornwall, 1992; Landed & Wimmer, 2008; Manis, Doi & Badha, 2000; McBride-Chang & Manis, 1996; Wimmer, 1993; Wolf & Bowers, 1999).

Some aumors consider RAN as an indicator of a mechanism mat is essential for automatic processing of orthographic (rather than phonological) representations in memory, i.e. slow letter identification would make it difficult for children to acquire the sensitivity for frequently occurring orthographic patterns (e.g. Bowers & Wolf, 1993; Wolf, Bowers & Biddle, 2000; for a more detailed discussion see Kirby, Parilla & Pfeiffer, 2003 or Georgiou, Parilla, Kirby & Stephenson, 2008; for a recent finding challenging this view see Moll, Fussenegger, Willburger & Landerl, 2009). Other aumors (e.g. Spring & Davis, 1988) assume that RAN might reflect the ability to coordinate multiple concurrent processes (i.e. visual, linguistic and articulatory) or general cognitive processing speed (e.g. Kail, Hall & Caskey, 1999).

For the predictive strength of early PA and RAN measures at least three methodological aspects are critical (Verhagen, Aamoutse & van Leeuwe, 2008): The manner in which PA and RAN are measured, the implemented reading measures (i. …