(i) previously did not require a supplemental sample and the recomputed coefficient of variation does not exceed the survey limit, no additional samples are required;

(ii) previously did not require a supplemental sample and the recomputed coefficient of variation exceeds the survey limit, additional samples are required;

(iii) previously required a supplemental sample and the recomputed coefficient of variation exceeds the survey limit, new additional samples are required; or

(iv) previously required a supplemental sample and the recomputed coefficient of variation does not exceed the survey limit, no additional samples are required.

(i) During the computation of the market value ratio estimate pursuant to this Subpart, the contribution to the coefficient of variation of the estimate, or CV, is also estimated for each value interval.

(ii) For the computation of the coefficient of variation for a class in a special assessing unit, each “U” interval is assumed to have a relative sampling variance proportional to its corresponding special assessing unit class relative variance. For the computation of the coefficient of variation for a class in a nonspecial assessing unit or the Amityville or Wyandanch School District in the Town of Babylon, each “U” interval is assumed to have a relative sampling variance proportional to its corresponding municipal or school district relative variance.

(iii) A coefficient of variation is not computed for the utility class. “S” intervals for the utility class are assumed to make no contribution to the sampling error since the intervals are not assigned their sample parcels by random selection methods.

(iv) “T” intervals make no contribution to the sampling error.

(v) All other “S” intervals contribute to the sampling error.

(vi) The estimated coefficient of variation for each estimate is obtained by expressing the estimated standard error as a percentage of the ratio estimate.

(i) If necessary, a supplemental sample is determined and the estimated coefficients of variation are tested in a multistep process. The order of the steps is as follows:

(ii) After the first step, the testing of the municipal estimated coefficient of variation, any hypothetical reduction to the CV attributed to any supplemental sample required in the first step or subsequent steps is assumed to have actually occurred prior to the test for the next step.

(iii) At each step, the estimated coefficient of variation is compared to its corresponding CV limit shown herein.

(iv) If the CV is greater than the appropriate CV limit, the sample size of the “S” interval contributing most to the CV is increased by one, the value interval contribution and CV are reduced hypothetically in accordance with sampling theory, and the adjusted CV is further tested against the CV limit. This process continues until a minimum number of supplemental samples have been assigned or the adjusted CV no longer exceeds the appropriate CV limit. If the initial CV estimate does not exceed the appropriate CV limit, no supplemental sample is assigned for the step. The minimum supplemental sample assigned for the municipal level estimate step is six, and for the village, borough, Amityville and Wyandanch School Districts in the Town of Babylon, school district in a school district segment survey assessing unit, and school district designated in former 9 NYCRR section 186-6.1, as repealed in 1998 and set forth in section 1230 of the Real Property Tax Law in a special or homestead assessing unit level estimate steps the minimum supplemental sample assigned is four. There are no minimums for the other estimates tested.

(v) The CV limits for New York City are:

(vi) For a class at the special assessing unit level, the CV limit is five percent.

(vii) The CV limits for other assessing units are: