SEVERE DOES NOT ALWAYS IMPLY SIGNIFICANT - BIAS OF A REGRESSION DISCREPANCY MODEL

The Regression Discrepancy Model (RDM) was originally offered as a method to increase accuracy in assessing severe discrepancy by accounting for regression effects and measurement error. However, demonstration that the RDM realizes its goal of accurate, nonbiased assessment of IQ-achievement discrepancies has not been provided. The RDM's mathematical equations are examined, particularly the initial equation evaluating significant difference between IQ and achievement scores. Results from the initial equation are shown to be related to IQ levels and to bias the detection of severe discrepancy at lower IQ levels. Methods to minimize or remove this bias are proposed.