Evaluating the GAPS test as a screener for language impairment in young children
Hannah Nash1, Ruth Leavett1, Helen Childs1
1Department of Psychology, University of York, UK.
Background:
The early identification of children is one of five themes identified by the Bercow review of 2008. The review also notes that there is a wide range in the methods used to identify children and it goes on to recommend that there needs to be a more systematic approach. One such approach would be to screen children before, or shortly after, school entry. The GAPS test has been designed as a screening tool to identify young children with language impairment and is reported to be of value in identifying children with language difficulties. However, the test has previously only been evaluated by its authors and the sensitivity of the test for identifying children from an unselected sample has not been evaluated.
Aims:
This study evaluated the ability of the GAPS test to identify language-impaired children in an unselected sample. In addition, the effect of tester status (a trained researcher and a teaching assistant) was investigated.
Methods & Procedures:
A total of 106 children aged 3-6 years completed the GAPS test, the Early Repetition Battery (ERB) and the core language scales from the Clinical Evaluation of Language Fundamentals-Preschool 2 with a trained researcher. Half the children completed the GAPS test a second time with a teaching assistant.
Outcomes & Results:
There was a significant effect of tester only for the non-word repetition subtest of the GAPS test in the nursery age group; the teaching assistants awarded higher scores than trained researchers. Of the 106 children, ten were language impaired according to the CELF-Preschool 2 core language score. The GAPS test identified two of these children at the 10th percentile cut-off, resulting in a low sensitivity estimate of 20%. However, the GAPS test only identified four of the 96 remaining unimpaired children resulting in a high specificity value of 96%. These values were similar when the 15th percentile cut-off was used and when parental concern or a family history of reading difficulties were used as the criterion measure.
Conclusions & Implications:
These data show that although the GAPS test can be used by a range of people who work with young children, it is not a sensitive screener for language impairment when used by trained researchers.
