Table 13. Comparisons of sensitivity and specificity measures of tests across three studies.

Study (classifying SLI vs. controls)	Criterion	Phonology:		Grammar:		Grammar
		Non-word repetition		Sentence Repetition		Elicitation of Past tense Marking
		Sensitivity	Specificity	Sensitivity	Specificity	Sensitivity	Specificity
Conti-Ramsden, 2003	16th	59%	100%			52%	100%
Botting & Conti-Ramsden, 2003	16th	79%	87%	90%	85%	89%	89%
GAPS sub-tests	15th	91%	93%	91%	100%

Table Legend: Comparison of the different test content across similar and different paradigms (non-word repetition or sentence repetition) or elicited production taken from two previous studies and this study and their resulting sensitivity and specificity measures. Here the two GAPS subtest are compared separately; one under grammar and one under phonology. For measures of phonology the GAPS-phonology subtest was the most sensitive across the studies. However, the NWRep was more specific in Conti-Ramsden's 2003 study. However, as in this study, Conti-Ramsden found that 40% of the control children were incorrectly identified as “impaired”. The higher specificity of the GAPS Grammar subtest also differentiates it from the CELF-RS.