Table 4.

Technical Adequacy of Cross-Battery Method (XBA)

Cognitive Abilitya	Low Construct Cutoffb	Base Rate %c	Sensit.d	Specif.e	PPVf	NPVg	E1h	E2i
Phonemic Awareness	−.75	5.4	.47	.96	.38	.97	94.6%	92.9%
	−1.00	6.5	.54	.97	.52	.97	93.5%	93.8%
	−1.25	6.8	.51	.96	.51	.96	93.2%	93.3%
RANj	−.75	7.0	.46	.94	.36	.96	93.0%	90.6%
	−1.00	8.6	.55	.95	.53	.96	91.4%	91.9%
	−1.25	9.1	.51	.95	.52	.95	90.9%	91.2%
Verbal	−.75	3.7	.46	.96	.33	.98	96.3%	94.6%
	−1.00	4.9	.51	.97	.48	.97	95.0%	94.9%
	−1.25	5.7	.44	.97	.49	.97	94.3%	94.2%
Memory	−.75	6.3	.47	.95	.38	.95	93.7%	91.9%
	−1.00	7.6	.54	.96	.53	.96	92.4%	93.0%
	−1.25	7.8	.51	.96	.51	.96	92.2%	92.4%

Notes:

^aCognitive Ability is the deficient variable for a given model.

^bLow Construct Cutoff is the z score cutoff used to define low ability at the latent level.

^cBase Rate % is percentage of cases in the sample selected for low achievement but of average mental ability classified as SLD at the latent level.

^dSensitivity is the proportion of true SLD cases identified as such by the method.

^eSpecificity is the proportion of true Not SLD cases identified as such by the method.

^fPPV is Positive Predictive Value or the probability of being truly SLD given a positive test result.

^gNPV is Negative Predictive Value or the probability of being truly Not SLD given a negative test result.

^hE1 is the efficiency or percent of correct classifications made without applying the test.

ⁱE2 is the efficiency or percent of correct classifications made when the test is used.

^jRAN = rapid automatic naming.