"In the February election, the Lake Washington School District bond measure reached a majority of yes votes but did not get the 60 percent supermajority needed to pass. As a result, the District commissioned a survey of district residents to explore concerns around the bond measure and attitudes about the school district. The good news is that those surveyed generally are impressed with the district. Residents cited concerns about the economy and taxes as the most likely causes for less support for the bond measure.“Given the cost of running a bond or levy election, it’s critical that we know what our community thinks,” noted Dr. Chip Kimball (left), superintendent. “What we heard from district residents is that now is not the time to ask for money for school construction, even though in general residents are very positive about the district. As a result, we will not run the bond measure again this year but will look carefully at the most cost-effective alternatives for dealing with the student population that will be coming in the next few years.”" Read full press release
###
According to an email by Kathryn Reith, LWSD Communications Director:
400 LWSD residents were surveyed at random. "Residents of the district were surveyed and we have many staff members who live in the district, so they may have been part of the survey but there is no way to tell. It was done by an independent research company. Some of the questions in the survey are ones that this group has used over many years with other districts in the state while others are very specific to this situation, of course."
Is this 400 resident survey of value if it is not statistically significant?. Almost 100,000 voters populate LWSD. Was the taxpayer money spent on this survey worth it?
By Bob Yoder
Photo by LWSD
Bob:
ReplyDeleteWith the sample of 400, the percentages are statistically valid to a margin of error of +/-5% at the 95% confidence level. That's pretty standard for a telephone survey of this type. More phone calls to narrow down the margin of error would have been very costly.
Kathryn Reith
Communications Director, Lake Washington School District
Thank you for clarifying, Kathryn; however in my college science classes 'statistically valid' was never in our vocabulary.
ReplyDeleteA statistical survey classified as "valid" sounds subjective and judgemental rather than objective and scientific.
I wonder what others have to say about this?
Bob, your knowledge needs updating. Go to Bing or Google and you'll find more than you'll need to know about statistically valid methodology. For simplicity, here are 2 links for you.
ReplyDeleteIf your statistics are up to is, read this
http://www.variation.com/techlib/as-7.html
A more layman-friendly explanation is here
http://getgood.typepad.com/getgood_strategic_marketi/2006/01/the_vagaries_of.html
You may be right. I didn't look at your links, but found a sample size calculator (below), ran the numbers and came out with a 393 sample requirement,based on assumptions made. It's still somewhat foggy, but I think I'm seeing some light on this and thank you for alerting me to your research.
ReplyDeletehttp://www.marketingmo.com/how-to-articles/marketing-metrics/how-to-determine-an-effective-sample-size/