English abstract
The purpose of this study was to examine the diagnostic accuracy of three Early Numeracy Curriculumn Based Measurements (EN-CBM). These measures were designed to asses three different skills of number knowledge including Number Identification, Missing Number and Quantity Discrimination. The measures were administered to 149 first grade students from Sharqia South Governorate. Results of internal consistency reliability indicated that Cronbach's Alpha correlation coefficient ranged between 0.81 and 0.97. In addition, stability equivalence reliability showed that Pearson correlation reliability ranged between 0.72 and 0.97 for the three measures. The measures' validity was estimated by sampling validity through preparing specification tables. Face validity exhibited 100% of a panel of experts' agreement about regarding the measures. Receiver Operating Characteristic curve (ROC) analysis was applied to examine diagnostic accuracy for the EN-CBM, using the Learning Disabilities Diagnostic Inventory (LDDI) as the outcome measure. The diagnostic accuracy examined the extent to which the EN-CBM can predict students' performance at or below thie 25th percentile. Area under the curve ranged from 0.81 to 0.83, sensitivity ranged from 0.80 to 0.91, specificity ranged from 0.65 to 0.70, Positive Predictive Value PPV ranged from 0.47 to 0.57, Negative Predictive Value NPV ranged from 0.89 to 0.91 and diagnostic accuracy ranged from 0.69 to 0.74. Overall, results showed that EN-CBM can predict student performance at early numeracy skills
