Chapter 1. Introduction
1.5. Evidence-Based Practice Recommendations for Assessment
A lack of consensus exists in relation to assessment and diagnosis of language disorder in children (Reilly et al., 2014), with literature noting variations across political jurisdictions and service agencies with regards to policy related to assessment practice (Ireland, Hall-Mills, & Millikin, 2013; Spaulding, Szulga, & Figueroa, 2012). Although no detailed formal guidelines exist to guide language assessment practice at a national or international level, professional literature does outline a number of evidence-based
assessment recommendations for school-aged children with language disorder (Bishop et al., 2016). These recommendations are discussed in this section of the thesis.
1.5.1. Using assessments that target different contexts and tasks. Evidence-based
practice recommends that SLPs should use a variety of language measures and assessment procedures when evaluating the language abilities of a school-aged child (American Speech- Language-Hearing Association, 2000; Bishop et al., 2016; Ebert & Pham, 2017; Speech Pathology Australia, 2011a, 2015). Assessment findings may vary depending on the context in which the child’s skills are assessed or the types of tasks used within the assessment (Harlaar, DeThorne, Smith, Betancourt, & Petrill, 2016). For example, a number of studies
have identified that results from different language measures conducted in a structured “test- taking” context correlate more highly with each other than they do with results from less structured “language sampling” procedures, suggesting that language performance may be influenced more so by the context than the actual skills being assessed (Dethorne, Johnson, & Loeb, 2005; Harlaar et al., 2016; Ukrainetz & Blomquist, 2002). Similarly, another study by Lennox, Westerveld, and Trembath (2018) identified that performance on a sentence level language measure (Renfrew Action Picture Test; Renfrew, 2010) was not highly predictive of performance on narrative retell task (Profile of Oral Narrative Ability; Westerveld, Gillon, & Boyd, 2012). Therefore, use of assessments that target different contexts and use different types of tasks are needed to fully describe a child’s strengths and weaknesses (Bishop et al., 2016). In addition, given that currently available comprehensive norm-referenced spoken language measures have psychometric limitations, assessing language abilities using a range of different assessments reduces the risk of missing important information on a child’s language abilities during diagnostic process (Eadie et al., 2014).
1.5.2. Selecting assessments that match the purpose of the assessment. Different
assessments are suited for different purposes. Language measures with set guidelines for administration and scoring are easily repeatable and relatively quick to administer and analyse (Kaderavek, 2015). However, a weakness of these types of assessments is that tasks are typically not representative of the child’s natural communicative environments and are, thus, not suitable for judging functional performance in everyday environments (Kaderavek, 2015; Pavelko, Owens, Ireland, & Hahs-Vaughn, 2016). Many language measures with set guidelines for administration and scoring are designed to cover a comprehensive array of skills in one language measure; however, in doing so may fail to assess each skill in enough depth to identify suitable intervention goals or effectively detect changes over time (Ebert & Scott, 2014).
Assessment procedures without set guidelines for administration and scoring allow for a child’s abilities to be evaluated in naturalistic contexts and are thus well-suited for
developing intervention goals that target improved performance in everyday activities (Kaderavek, 2015; Yont, Hewitt, & Miccio, 2002). Specific skills may also be examined in greater depth than language measures with set guidelines for administration and scoring might allow. However, assessment procedures without set guidelines for administration and scoring cannot be accurately repeated for the purposes of empirically comparing the
performance of different children or comparing the performance of the same child across time (Watson & Pennington, 2015). By using a range of both language measures and
assessment procedures, SLPs will have appropriate data to use for the different purposes for which assessment data are often required (Fulcher-Rood et al., 2018).
1.5.3. Selecting appropriate assessments for children from CALD backgrounds.
Assessments with norm-referenced data should only be used when the normative sample population matches the child’s background (Kaderavek, 2015). When a child’s demographics or background is different to the normative population, an accurate comparison of a child’s performance in relation to peers is not possible. For this reason, assessments with normative data from monolingual English-speaking children are not appropriate for children who have culturally and linguistically diverse (CALD) backgrounds, including children who are bilingual or learning English as a second language (Arias & Friberg, 2015; Caesar & Kohler, 2007; Pearce & Williams, 2013). For example, one study found the Australian version of the Clinical Evaluation of Language Fundamentals – 4th Edition (CELF-4; Wiig, Semel, &
Secord, 2004) to be biased against Aboriginal English language forms and, therefore, likely to identify language difference as being a language disorder in children with Aboriginal backgrounds (Pearce & Williams, 2013). Instead of assessments with norm-referenced data, evidence-based practice guidelines recommend that SLPs should use other types of
assessments to diagnose language disorder in children from CALD backgrounds (Caesar & Kohler, 2007). One recommended alternative is dynamic assessment, which investigates a child’s learning potential by examining the child’s responses to explicit teaching (Peña, Gillam, & Bedore, 2014). Findings from previous studies have identified that dynamic assessment has clinical utility when differentiating language difference from language disorder in children from CALD backgrounds (Peña, Gillam, & Bedore, 2014; Peña et al., 2006). In addition, dynamic assessment also has usefulness for identifying a child’s potential for future language growth (Binger, Kent-Walsh, & King, 2017; Olswang & Bain, 1996).
1.5.4. Collecting data from parents and teachers as an assessment method.
Evidence-based practice guidelines recommend that SLPs should collect data from parents or teachers when conducting an assessment of a child’s language abilities (Bishop et al., 2016). Information from parents and caregivers provide unique information on a child’s functional abilities in everyday contexts that supplements results of assessment conducted by SLPs. For example, in a study by Bishop and McDonald (2009) found that identification of language disorder was most accurate when scores from language measures conducted by an SLP were combined with parental report. Similarly, another study measuring intervention outcomes identified that parent report captured information on social abilities that may not be identified through other language measures (Thomas-Stonell, Washington, Oddson, Robertson, & Rosenbaum, 2013). Teacher ratings of a child’s language ability also provide valuable diagnostic information, particularly for children from CALD backgrounds, whose abilities may not be accurately assessed through norm-referenced measures (Pearce & Williams, 2013). Collecting information from parents and teachers during the assessment process is also important for developing intervention goals that are aligned with the concerns of families or targeted at addressing educational needs (Crais, Roy, & Free, 2006; Dockrell & Lindsay, 1998).
1.5.5. Using ICTs as an assessment method. Research has identified that language
assessment for school-aged children may be effectively conducted using ICTs (i.e. telehealth; Edwards, Stredler-Brown, & Houston, 2012) A study by Waite, Theodoros, Russell, and Cahill (2010b) identified no significant differences between ICT or face-to-face delivery of a norm-referenced language measure for children. In similar studies, a high level of agreement was also obtained between ICT and face-to-face delivery of a range of other language and literacy assessments for children, including measures of phonemic awareness, spelling and reading (Ciccia, Whitford, Krumm, & McNeal, 2011; Waite, Theodoros, Russell, & Cahill, 2010a). Given that many children and families experience lengthy travel times to access services (O'Callaghan, McAllister, & Wilson, 2005; Ruggero, McCabe, Ballard, & Munro, 2012), use of ICTs for conducting language assessment is recommended for improving service accessibility and reducing travel time and costs (Edwards et al., 2012; Mashima & Doarn, 2009).