About a Survey 

I promised last week that I would write about a recent survey that seemed to rely on a skewed, unrepresentative sample. The survey and some printed accounts of it have been available, but the recording of the 6.6.16 meeting at which the results were initially presented does not seem to be available online for readers. 

I’ll say now that the conclusions of the survey seem to be right (that a school referendum would likely pass) but that the sample the survey uses to reach that conclusion is strangely unrepresentative of the community it purports to describe. That matters because the right conclusion with the wrong data is little more than guesswork.  Surveys are not meant to be paid guesswork.

It also matters because although the primary conclusion may be right despite a weak sampling of the community, other inferences drawn from the survey may not be similarly accurate. 

Looking at actual election data, actual demographic data, and past referendum results will produce a better assessment than relying on the recent community survey’s data sample.  

It seems fair to the survey authors (an outside vendor), however, to include for readers their full, 6.6.16 meeting remarks.  If they’re published online soon I’ll include them when publishing my assessment; if they’re not available I’ll go ahead with the post with the published information that is available.

(News accounts of the survey only reveal that those accounts’ authors either don’t understand or don’t care how unrepresentative the survey sample is of the electorate that would be considering a referendum, or even the community as it is now.)

It’s worth coming to the right conclusion with the right data; anything less is less than this community deserves. 

More to come.