The development and validation of the Short Language Measure (SLaM): A brief measure of general language ability for children in their first year at school.

BACKGROUND: There is no sufficiently accurate short-language measure that could be used by speech-language pathologists, teachers or paraprofessionals to screen young school-aged children to identify those requiring in-depth language evaluations. This may be due to poor development of the available measures, which have omitted crucial test development steps. Applying more stringent development procedures could result in a measure with sufficient accuracy. AIMS: To create and validate a short-language measure that has acceptable accuracy, validity and reliability, and which can be used to identify children who require further assessment and/or referral to speech-language services. METHODS & PROCEDURES: The study consisted of two phases. In Phase 1 (measure creation), 56 children were assessed with 160 direction-following and sentence-recall test items and a reference measure, the Clinical Evaluation of Language Fundamentals-Fourth Edition (CELF-4). Items were then examined for their individual characteristics (validity, reliability, difficulty and discrimination) via item analysis and the highest quality items were selected to form the Short Language Measure (SLaM). In Phase 2 (measure validation), 126 children were assessed with the SLaM and the reference measure (CELF-4) to determine SLaM's accuracy, validity and reliability. OUTCOMES & RESULTS: A total of 40 test items were selected to form SLaM in Phase 1. Findings from Phase 2 indicated that SLaM had an accuracy of 94% (sensitivity = 94%, specificity = 93%), validity of 0.89 and reliability of 0.93. These values remained relatively consistent across both phases. CONCLUSIONS & IMPLICATIONS: The results indicated that SLaM has excellent psychometric properties. It can be used to identify children who need further evaluation by a speech-language pathologist. What this paper adds What is already known on this subject Prior research suggests that combining a direction-following and a sentence-recall task has sufficient discrimination accuracy and agreement with an omnibus language measure. Trialling a large set of direction-following and sentence-recall test items to select those with the highest individual characteristics could result in an effective short-language measure. What this paper adds to existing knowledge A short-language measure (SLaM) was created and validated on two independent samples of children. Items with the highest validities, reliabilities and discrimination capacities were selected to form SLaM. This procedure resulted in a measure with high validity and reliability that exceeded the criterion for adequate discrimination accuracy. What are the potential or actual clinical implications of this work? SLaM is an effective measure that can accurately identify children who require detailed evaluations by speech-language pathologists.

Investigation of the language tasks to include in a short-language measure for children in the early school years.

BACKGROUND: The inaccurate estimation of language difficulties by teachers suggests the benefit of a short-language measure that could be used to support their decisions about who requires referral to a speech-language therapist. While the literature indicates the potential for the development of a short-language measure, evidence is lacking about which combination of language tasks it should include. AIMS: To understand the number and nature of components/language tasks that should be included in a short-language measure for children in the early school years. METHODS & PROCEDURES: Eight language tasks were administered to participants of the Early Language in Victoria Study (ELVS) at ages 5 (n = 995) and 7 (n = 1217). These included six language tasks measured by an omnibus language measure (which comprised a direction-following, morphological-completion, sentence-recall, sentence-formation, syntactic-understanding and word-association task) and a non-word repetition and a receptive vocabulary task, measured by two task-specific language measures. Scores were analyzed using principal component analysis (PCA), the Bland and Altman method, and receiver operating characteristic (ROC) curve analysis. OUTCOMES & RESULTS: PCA revealed one main component of language that was assessed by all language tasks. The most effective combination of two tasks that measured this component was a direction-following and a sentence-recall task. It showed the greatest agreement with an omnibus language measure and exceeded the criterion for good discriminant accuracy (sensitivity = 94%, specificity = 91%, accuracy = 91%, at 1 SD (standard deviation) below the mean). CONCLUSIONS & IMPLICATIONS: Findings support the combination of a direction-following and a sentence-recall task to assess language ability effectively in the early school years. The results could justify the future production of a novel short-language measure comprising a direction-following and a sentence-recall task to use as a screening tool in schools and to assess language ability in research participants.