Severe discrepancy regression worksheet
Computes the ability–achievement discrepancy for SLD eligibility under 5 CCR §3030(b)(10) using the regression method, and shows the arithmetic behind every result. Scores stay in this browser — nothing is sent anywhere or saved.
Before you use this worksheet. A severe discrepancy is one condition of SLD eligibility, not the decision. No single score, test, or procedure may be the sole criterion for the team's eligibility decision, and a computed discrepancy counts only when it is corroborated by other assessment data.
1Student and evaluation
Identifiers are printed on the worksheet you keep. They are never transmitted or stored by this tool.
2Intellectual ability
Enter the standard score on the mean-100, SD-15 scale. Use the composite that is the best measure of this student's ability.
Mean 100, SD 15. Convert first if the measure reports another scale.
3Correlation and measurement error
Pick the cognitive measure and r fills in from the district table. Both values print on the worksheet with their source.
Choose the composite that is the best measure of this student’s ability — the same one entered in step 2. WISC–V, DAS–2, and WAIS composites are all on the district table.
Filled from the selection. Override it if the district has adopted a different value for this test pair.
Stamped automatically from the selection above. This is what makes the value defensible on review — edit it if the district adopts its own table.
The regulation permits the size of a severe discrepancy to be decreased by one standard error of measurement, up to a maximum of 4 standard score points. It does not require it. The examiner decides, and the worksheet prints which was done.
From the WIAT–4 manual, for the composite or subtest being compared and this student’s age band.
4Achievement areas
One row per area assessed. The eight areas are the ones named in the regulation; the WIAT–4 measure fills in from the district’s designated list.
5Required considerations
A discrepancy alone does not establish eligibility. Confirm each of these before the team meets.
6How this is computed
“…a severe discrepancy is demonstrated by: first, converting into common standard scores, using a mean of 100 and standard deviation of 15, the achievement test score and the intellectual ability test score to be compared; second, computing the difference between these common standard scores; and third, comparing this computed difference to the standard criterion which is the product of 1.5 multiplied by the standard deviation of the distribution of computed differences of students taking these achievement and ability tests. A computed difference which equals or exceeds this standard criterion, adjusted by one standard error of measurement, the adjustment not to exceed 4 common standard score points, indicates a severe discrepancy when such discrepancy is corroborated by other assessment data…”
5 CCR §3030(b)(10)(B). Quoted identically in the Riverside County SELPA and El Dorado County SELPA procedural guides.
The regression method, step by step. The regression method predicts the achievement score expected for this student's ability, then asks whether the obtained score falls far enough below that prediction. It corrects for regression toward the mean, which a plain subtraction does not.
1. Predicted achievement = 100 + r × (ability − 100)
2. Standard error of estimate (SEe) = 15 × √(1 − r²)
3. Standard criterion = 1.5 × SEe
4. Adjusted criterion = standard criterion − min(SEM, 4), when the SEM reduction is applied
5. A severe discrepancy is indicated when (predicted − obtained) ≥ the adjusted criterion
This is the same arithmetic published as a lookup table by the New Mexico Public Education Department, SLD Severe Discrepancy Regression Table (“Size of Discrepancy = 1.5 SD × SEe”), which converts the same formula into a lookup table at r = .7, .6, and .5. That table rounds its printed cells to whole scores; this worksheet compares the unrounded values, so a borderline case can differ from the table by a point. For California-specific criterion and SEM tables by test pair, see Helm, D. P., & Johnson, D. E. (1998). Updated tables needed to determine a severe discrepancy under California's Title 5 regulations. The California School Psychologist, 3, 19–33.
The standard error of measurement is optional. The regulation permits the size of a severe discrepancy to be decreased by one standard error of measurement, up to a maximum of 4 standard score points — it does not require it. The worksheet defaults to not applied, so using it is a deliberate choice the examiner makes and the printed worksheet records. When it is applied, the reduction is capped at 4 points.
The second figure shown on each row, simple difference, is the regulation's text read literally — ability minus achievement, against 1.5 × 15√(2 − 2r). It is shown for reference. The headline result is the regression method.