We use a sample design for selecting sites for the evaluation sample that takes such skewness into account.

Cleared for reuse. Confirm the cited sources still say this before you paste it.

Status

• Review verdict: approve
• Reusable as written: yes
• Issues found: generic_usable
• Reviewer note: The text describes a generic statistical sampling methodology without specific client names, dates, figures, or confidential data. It is reusable for any evaluation proposal requiring site selection.

Text

We use a sample design for selecting sites for the evaluation sample that takes such skewness into account. With this approach, a number of sites with large savings are selected for the sample with certainty, and a random sample is taken of the remaining sites. To further improve the precision, non-certainty sites are selected for the sample through systematic random sampling. That is, a random sample of sites remaining after the certainty sites have been selected is selected by ordering them according to the magnitude of their savings and using systematic random sampling. Sampling systematically from a list that is ordered according to the magnitude of savings ensures that any sample selected will have some units with high savings, some with moderate savings, and some with low savings. Samples cannot result that have concentrations of sites with atypically high savings or atypically low savings. Using this procedure essentially reduces the coefficient of variation for the non-certainty sites that are sampled.

Applicability

• Section: sampling_analysis
• Scope: mixed_context
• Observed documents: 3
• Years: 2011 to 2012
• Issuers: MISSING
• Program types: MISSING
• Jurisdictions: Oregon

Sources

• [F020409, paragraph:330]
• [F021446, paragraph:166]
• [F021685, paragraph:153]