10 C.F.R. § 429.28
(a) Sampling plan for selection of units for testing.
(2) For each basic model of faucet, a sample of sufficient size shall be randomly selected and tested to ensure that any represented value of water consumption of a basic model for which consumers favor lower values shall be no less than the higher of the higher of:
(i) The mean of the sample, where:

Or,
(ii) The upper 95 percent confidence limit (UCL) of the true mean divided by 1.05, where:

(b) Certification reports.
[76 FR 12451, Mar. 7, 2011; 76 FR 24771, May 2, 2011, as amended at 78 FR 62985, Oct. 23, 2013]