September 05, 2007
County Council authorizes statistical sampling
The King County Council yesterday voted unanimously to authorize the use of the state's statistical sampling process on county initiative petitions. Good. This will save considerable time and tax dollars on future initiative petitions.
The legislation was introduced by Reagan Dunn and co-sponsored by Jane Hague, Kathy Lambert, Bob Ferguson, Julia Patterson and Pete von Reichbauer.
Posted by Stefan Sharkansky at September 05, 2007
12:31 PM | Email This
How can statistical sampling accurately determine how many valid signatures there would be after throwing out the signatures of everyone who seems to have signed more than once? The number of duplicates should be fairly small, so the sample size necessary to determine it would be larger than would otherwise be required.
Not that there's any legitimate reason for throwing out all the signatures (rather than all-but-one) for anyone who seems to have signed twice. If there's a real signature for Fred Q. Johnson as well as a fake one, one and only one should count. I see no reason the fake signature should cancel out the real one, though.
Nonetheless, since KC has the policy of throwing out people's signatures in such cases, I'm curious how they plan to do that via sampling.
2. Supercat, I believe King County Elections said they throw out all-but-one as you suggest. Nonetheless you raise a good point: they would need to do some fairly complex math to project the number of duplicates from a sample. I wonder whether they'll do that?