| Four screening methods were compared to determine the extent to which each method accurately identified students who exhibited math and reading problems severe enough to warrant a full psycho-educational assessment for special education eligibility determination. Methods examined were teacher referral, two subtests from the Comprehensive Inventory of Basic Skills, Revised, the Developmental Reading Assessment, and a new screening measure called "Problem Validation Screening" (PVS). PVS consisted of three components: classwide curriculum-based measurement probes in math and reading, performance/skill deficit assessment, and a brief instructional session. All students enrolled in first and second grade classes at a participating school were exposed to each of the four screening measures. Students who met criteria as potentially exhibiting a serious problem in reading or math were then exposed to a second measure, termed the "Criterion Assessment" along with two traditional measures of student achievement (i.e., Woodcock-Johnson Psychoeducational Battery, Revised and Iowa Test of Basic Skills). The Criterion Assessment consisted of individually administered curriculum-based assessment and individual intervention. Sensitivity, specificity, as well as positive and negative predictive power were calculated for each screening measure using the Criterion Assessment as the outcome standard. The predictive accuracy estimates of each screening method were compared in classrooms where base rates of academic problems were high and classrooms where base rates of academic problems were low. In all cases, PVS achieved the highest predictive accuracy estimates of all the screening measures regardless of the sample base rates. Implications of the study for practice are discussed along with the importance of measuring both performance levels and trend when evaluating student performance. |
