OK, I’ll have to blow the dust off of it and read through the yellowed pages :).
Not sure self-selecting matters too much in this case (and I’m pretty sure my textbook says that self-selecting is bad). I didn’t feel the way the survey was presented led to a bias.
Anyway, just fishing for opinions. I still agree with Ian’s conclusions.
On 09/12/2013 10:39 AM, Doug Schaefer wrote:
Which is actually another point, the response size is pretty small. I’m not sure we can conclude anything based on that.
Check your old statistics textbook. The total sample size was well over 2000 people. That is definitely a statistically significant sample size. Of course the sample was self-selecting so there will be bias. So
please stop saying that the sample was too small.
If we assume a population of 9 million people who use Eclipse, a confidence level of 99%, and a confidence interval of 3%, we only needed 1849 responses.
http://www.surveysystem.com/sscalc.htm
--
Mike Milinkovich
mike.milinkovich@xxxxxxxxxxx
+1.613.220.3223
