ACQ Vol 13 no 3 2011

assessment principles address the potentially confounding aspects of standard forms of assessment (e.g., culturally specific question–answer routines). That is, CALD children who, for example, are not exposed to the direct nature of western speech pathology style questioning at home, might be misidentified as language impaired on the basis of responses that represent cultural difference rather than language difficulty. Dynamic assessments incorporate a learning component into the testing situation in preference to static assessment administration. The learner’s responsiveness to teaching is assessed. Test-teach-retest procedures have been identified as the most suited dynamic approach to SP assessment and intervention (Guttiérrez-Clellen & Peña, 2001). Such approaches, however, are limited to the diagnostics of learning impairment and do not necessarily provide specific information concerning where language breakdown occurs (Gutiérrez-Clellen et al., 2006). For example, dynamic testing (a shortened version of dynamic assessment) has been shown by Chaffey, Bailey, and Vine (2003) to provide valid data regarding high learning potential in a sample of rural NSW Australian Aboriginal primary schoolers (grades 3–5). This form of testing proved to be a more sensitive measure compared to alternative static cognitive testing, highlighting the potential of dynamic testing in school assessments. More recently, Kramer, Mallett, Schneider, and Hayward (2009) investigated the use of dynamic language assessments to assess narrative abilities of First Nations, grade 3 students on the Samson Cree Reserve, Alberta, Canada. The authors used the Dynamic Assessment Intervention tool (DAI; Miller, Gillam, & Peña, 2001) that was designed to minimise social and cultural bias when assessing language development with CALD children. The mediated test-teach-retest method was employed to test oral narrative constructions from wordless storybooks. Samples were scored according to content (e.g., establishment of time and place) and results showed that the DAI accurately differentiated most typical language learners from those learners with possible language- learning difficulties. Although Kramer et al. (2009) discussed the universality of the storytelling, the authors did not examine the cultural validity of the criteria used for scoring the stories. The cultural validity of scoring needs to be considered in light of cultural variability. That is, certain semantic features might have a different significance according to linguistic and/or cultural membership. This idea is based on the linguistic relativity hypothesis which suggests that perception is limited by the language in which we think and speak. For example, when telling a story, speakers of language X might preferentially refer to the place of an event over time of the same event, whereas speakers of language Y might consider the place far less important than the time. This does not limit the usefulness of dynamic assessment, but does remind users of the impact culture and language can have on interpretation of assessment results. Novel linguistic stimuli approach A proposed alternative method of limiting cultural and linguistic biases in language testing is to use novel stimuli in assessments. Non-word repetition tasks have been used to access verbal working memory since with careful construction, stimuli are not dependent on a participant’s lexicon (Gathercole, 1995). Stimuli are however dependent on phonological familiarity and thus must be constructed

according to the phonotactics of the target language. The evidence for their reliability with English-Spanish bilingual speakers in the United States is not yet established. Ellis Weismer et al. (2000) found supporting evidence whereas Guttiérrez-Clellen and Simon-Cereijido (2010) concluded that if this type of testing is to be completed, both languages need to be assessed and the testing should not be used to make diagnoses in isolation. Speech Pathology Australia (2009) similarly recommends the assessment of both/all of a CALD child’s spoken languages. The successes or shortcomings of using non-word stimuli with English-Spanish bilinguals compared with Indigenous Australian populations cannot be drawn without complication. For example, the inherently formal nature of the non-word repetition assessment and its non-meaningful stimuli (Gould, 2008b) suggests that in an Indigenous Australian environment, performance is potentially confounded by contextual cultural bias. A variation of formal non-word repetition tests was therefore trialled when assessing language development in an Australian Aboriginal community (Gould, 2008c). Gould (2008c) describes how she overcame cultural barriers by designing a non-word repetition task for use in the aforementioned longitudinal research project assessing language development of AE speakers. The trialled assessment is based on the familiar speech pathology subtest of the Queensland University Inventory of Literacy (QUIL; Dodd et al., 1996) and the Sutherland Phonological Awareness Test – Revised (Neilson, 2003). It is an elegantly designed adaptation of a non-word test involving the use of 18 phonotactically AE- relevant non-words (see Gould, 2008c for a full description of testing methodology). It differs from other non-word tests; while it requires the child to repeat the non-word, repetitions are elicited during a play-based activity rather than during a formal standardised repetition task. Overall, Gould (2008c) shows that the culturally sensitive administration of a culturally appropriate assessment tool helps to: identify contributing reasons for literacy development difficulties; give qualitative information as to the nature and severity of difficulties; highlight abilities which had not been considered or had been ruled out by formal testing; and identify the need for a hearing assessment. Clearly this culturally appropriate format of assessment contributes greatly to an overall picture of a child which is potentially more accurate than that drawn from formal, culturally biased assessments. At this stage, results of such a non-standard assessment are unable to be compared with norms. Gould (2008c) suggests that in the absence of norms, data analysis should be completed in conjunction with Aboriginal educators/co-workers. When adapting a standardised test, translation of linguistic stimuli alone is not sufficient to ensure validity when assessing a CALD child’s communication abilities (Carter et al., 2005; Speech Pathology Australia, 2009). Gould (2008c) highlighted the need for cultural translation and adaptation on a number of levels including environmental context, test format, examinee/examiner relationship, recognition of different learning styles, and recognition of cultural differences such as “shame”. Gould (1999 cited in Gould, 2008b) also showed that without accounting for these differences when testing communication development of Australian Aboriginal children, standardised tests are likely to result in the over diagnosis of language impairment.

123

ACQ Volume 13, Number 3 2011

www.speechpathologyaustralia.org.au