Current through Register Vol. 54, No. 45, November 9, 2024
Section 8a.7 - Statistical estimation and softwareThe audit results shall be computed by projecting the audit findings identified in the sample, as adjusted for outliers as provided in § 8a.6(3) (relating to selection of sample) to the population, regardless of whether the sample is a statistical sample or a block sample.
(1) When the Department employs the block sampling method, the standard error cannot be estimated.(2) When the Department employs the statistical estimation method, a standard error of the estimate shall be computed from the sample observations adjusted for outliers as provided in § 8a.6(3) to indicate the reliability of the estimated average, total or ratio. The Department may use software that has been designed in accordance with accepted statistical practices. The formulas utilized by the software will be available for examination by the taxpayer.(3) Except as otherwise mutually agreed to by the Department and the taxpayer, the number of observations in the sample will be chosen so that the projected sample will, on the average, yield an estimated precision within 25% of the midpoint of a 90% two-sided confidence interval. In determining the size of the sample, the Department will use the sample size selection table in paragraph (4). The estimated precision of the sample selected may be less than or greater than 25%, depending upon the variability in the sample data. The standard error and estimated precision will be calculated and reviewed with the taxpayer. The sample size will be increased upon the request of the taxpayer. The process of increasing the sample size will be repeated until mutual agreement is reached between the taxpayer and the Department on an acceptable number of observations.(4) The following sample size selection table identifies estimated sample sizes required to produce estimates with specified precision: Sample Size Selection Table
| Precision | Confidence Interval (2 sided) | Estimated Coefficient of Variation (CV) |
| 0.25 | 0.50 | 0.65 | 0.75 | 1.00 | 1.50 | 2.00 | 3.00 | 5.00 |
| 5% | 90% Normal | 68 | 271 | 650 | 609 | 1,083 | 2,435 | 4,330 | 9,742 | 27,060 |
| 10% | Deviate | * | 68 | 115 | 153 | 271 | 609 | 1,082 | 2,435 | 6,765 |
| 15% | 1.645 | * | 31 | 51 | 68 | 121 | 271 | 481 | 1,082 | 3,007 |
| 20% | * | * | 29 | 39 | 68 | 152 | 271 | 609 | 1,691 |
| 25% | * | * | * | 25 | 44 | 97 | 173 | 390 | 1,082 |
| 30% | * | * | * | * | 31 | 68 | 120 | 271 | 752 |
| 35% | * | * | * | * | 23 | 50 | 88 | 199 | 552 |
| 40% | * | * | * | * | * | 38 | 68 | 152 | 423 |
* Fewer than 20 sample observations are required.