The first measure that was included in this battery of assessments was spelling. Spelling was selected as an indicator of participants’ word level literacy that would allow for
evaluation within the group testing format. Although spelling is a word encoding skill rather than decoding skill, spelling tasks require application of the same underlying skills as word decoding tasks. These include knowledge of phonology, orthography, and morphology (Wolf & Kennedy, 2003). Strong spelling ability requires the development of proficient
phonological, morphological, and orthographic awareness to create a strong mental representation of the word (Moats, 2009; Wolter & Apel, 2010).
More basic tasks of word decoding skill (such as word identification and non-word reading) were not included in the assessment for several reasons. Firstly, many typical word decoding tasks are administered individually and would not be suitable for inclusion in a large group format. More pertinently, word decoding tasks such as word identification and pseudo-word reading have been reported to demonstrate very little contribution to reading comprehension in adults in higher education (Jackson, 2005; Macaruso & Shankweiler, 2010). Furthermore, findings from these studies of adults with higher literacy skill demonstrated that spelling was a much stronger predictor of reading comprehension than word decoding assessments typically utilised with children (e.g., non-word and real-word reading). The target population for the current assessment was very similar to the studies identified above. It was therefore hypothesised that spelling would provide the most appropriate measure of word-level literacy skill.
3.1.1.2. Inferencing
Given the literature outlining a stronger contribution to reading comprehension from linguistic comprehension skills than word decoding skills amongst adults than children (e.g., Cromley & Azevedo, 2007; Landi, 2010; Mellard et al., 2010), it was hypothesised that a focus on higher level language skills within the assessment battery would be most
comprehension when the complexity and content of a text increases (Cartwright, 2009). Hence this measure was chosen as the linguistic comprehension measure from the SVR due to the target participants being engaged in higher education; an environment in which this shift in text complexity and content is exhibited. Furthermore, inferencing has been
demonstrated to allow for the differentiation to be made between highly skilled readers and less skilled readers, even within groups of participants engaged in higher education (Hannon & Daneman, 1998; Long et al., 1994). It was hypothesised that the inclusion of an
inferencing measure would be beneficial in describing the variance in reading comprehension amongst the higher education population, due to prior research indicating that even highly educated individuals experience difficulty with aspects of inferencing (Franks, 1998), particularly when the demands of written text are increased (Rapp et al., 2007).
3.1.1.3. Working Memory
A working memory measure was incorporated into the assessment battery, as it has also been found to differentiate between typical and less skilled adult readers (e.g., Hannon, 2012; Hatcher, Snowling, & Griffiths, 2002; Macaruso & Shankweiler, 2010). Furthermore, the ongoing debate in the literature about the role of working memory in reading
comprehension has called for additional research to develop a more comprehensive
understanding of its application to adults (Van Dyke et al., 2014). Working memory has been shown to be a significant predictor of reading comprehension in children (Cain, Oakhill, & Bryant, 2004; Seigneuric et al., 2000). Conversely, the research has also argued that deficits that have been attributed to working memory may instead occur as a consequence of
underlying language difficulties (Nation, 2005; Nation et al., 1999; Stothard & Hulme, 1992). These conflicting views have arisen due to working memory typically having been measured using language-based tasks. The task developed for utilisation in the current research was
comprised of numerical content, to limit the linguistic requirement of the task. Finally, due to the increased complexity, content, and number of texts experienced by individuals in higher education (Cogmena & Saracaloglub, 2009; Fidler & Everatt, 2012), it was hypothesised that working memory would be relied on more heavily in this context.
3.1.1.4. Knowledge of Language Structure
Four different aspects of metalinguistic knowledge were included in the knowledge of language structure component of the assessment battery. An increasing number of studies of metalinguistic ability have demonstrated low knowledge across several components,
particularly in the pre-service teacher population (e.g., Fielding-Barnsley, 2010; Moats, 1994; Spear-Swerling & Brucker, 2006; Washburn, Joshi, & Binks‐Cantrell, 2011). Knowledge and proficiency of the more intricate constructs of the English language (e.g. phonology,
morphology, and orthography) is crucial for teachers to enable them to provide effective literacy instruction in their professional practice (Moats, 2014).
Morpheme knowledge was included in the assessment battery as it has been demonstrated to be predictive of spelling ability (Masterson & Apel, 2007; Shankweiler, Lundquist, Dreyer, & Dickinson, 1996), and thus indirectly contribute to reading
comprehension. Phonological awareness has been reported as a strong predictor of reading comprehension (Al Otaiba, Kosanovich, & Torgesen, 2012; Gillon, 2004; Goswami, 2000), therefore phoneme knowledge (a phoneme identification task) was included as a component skill of phonological awareness. Further, it has been reported to be a key component within spelling and reading development (Goswami & Bryant, 1990; Stanovich, 1986). Orthotactic knowledge (e.g., knowledge about the spelling rules of the English language) has also been shown to be very important for spelling instruction (Moats, 2009; Snow, Griffin, & Burns, 2005), and may indirectly contribute to reading comprehension. All four of the metalinguistic
measures were included in the assessment battery not only to assess individuals’ knowledge in contributing to their own reading comprehension, but also to investigate their underlying knowledge of the skills required to provide effective explicit instruction in reading and spelling in their professional practice.
3.1.1.5. Reading Comprehension
Reading comprehension was included in the assessment battery as the outcome (independent) measure of reading. There are many standardised assessments available for, and normed upon, the school-age population (Cain & Oakhill, 2006a). In contrast, there is a paucity of standardised assessments designed specifically for the adult population (Kruidenier et al., 2010), and fewer still that are appropriate to use with adults of higher literacy skill (e.g., those engaged in higher education). An assessment that was standardised on the higher education population within the United Kingdom – the Adult Reading Test (ART; P. Brooks, Everatt, & Fidler, 2004) – was determined to be the most appropriate measure of reading comprehension for the target population. This assessment had previously been approved by the British Dyslexia Association, and argued to be comparable to the Passage Comprehension subtest of the Woodcock Reading Mastery Test (Woodcock, 1987) in earlier research (Fidler, 2009). Furthermore, the ART included multiple texts of the same reading level, thus allowing for inclusion of different texts of the same level to be utilised at pre- and post-assessment to negate practice effects. One text and the corresponding questions from two levels of the ART were included to provide a measure of reading comprehension within the proposed
assessment battery (outlined below).