Robin van Rijthoven

80 Chapter 4 knowledge of morphological and orthographic patterns further strengthens the quality of lexical representations. Indeed, it has been found that high-quality phonological, orthographic, and semantic representations lead to better spelling levels (Perfetti, 2007). Building up high-quality phonological and orthographic representations is demanding for all beginning spellers (as described by Cassar et al., 2005). However, for children with dyslexia, learning to spell words correctly is even more challenging (Lyon et al., 2003), due to a phonological deficit (Conrad, 2008; Göbel & Snowling, 2010). This deficit makes it difficult to learn and retrieve phonological information from long-term memory (Cassar et al., 2005), resulting in more phonological, morphological, and orthographic spelling errors as compared to typically developing children (Bourassa & Treiman, 2003; Bourassa et al., 2006; Cassar et al., 2005). The phonological deficit is an important predictive factor for spelling development (Wimmer & Mayringer, 2002). However, as proposed by the phonological-core variable differences model by Stanovich (1988), along with the core phonological deficit, other variables influence the spelling development as well. This model posits that all poor readers have a phonological deficit, but that additional cognitive factors are important for their reading and spelling skills. The triangular framework (Seidenberg & McClelland, 1989) highlights the importance of high-quality semantic representations for learning to read and spell. In line with this model, it can be assumed that variation in semantic abilities could partly compensate for the phonological deficit, especially since the semantic abilities of children with dyslexia are quite similar to those of typically developing children (Nation & Snowling, 1998). Phonological and semantic skills influence reading development from its earliest stages and influence each other as well (e.g., Laing & Hulme, 1999). However, Nation and Snowling (2004) showed that within the typically developing readers, broader language skills (also verbal production and comprehension) beyond phonology make important contributions to word recognition development. Prior research regarding reading indeed showed the benefits of well- developed semantic representations (e.g., Nation & Snowling, 2004; Torppa et al., 2010) and suggested a possible importance of semantic representations for spelling as well (Ouellette, 2010; Ouellette & Fraser, 2009). It remains, however, unclear to what extent variation in semantic ability impacts phonological versus orthographic spelling of children with dyslexia on top of phonological abilities. Analysis of spelling errors Only few studies addressed the distribution of different types of spelling errors (e.g., Bourassa & Treiman, 2003) and none predicted spelling errors from individual differences in cognitive factors. One way of addressing phonological and orthographic errors separately is to analyze and categorize spelling errors. Spelling error analyses and