Reply to thread

Message

<blockquote data-quote="Oofta" data-source="post: 9407909" data-attributes="member: 6801845">With a scale of only 1 thru 5, I just don't see the 1s or 5s having an outsized impact.&nbsp; It's not like we're looking at a dozen values that between 1 and 1,000.&nbsp; Taking the median when you have millions of responses does not mean you ignore all the results between 2 and 4 any more than if you had used the average.I also haven't seen a quote that said they used the median, although I could have missed it.&nbsp; Sometimes you want to use the median, sometimes you want to use the average for analysis, neither is preferable.&nbsp; Since I'm not a statistician I have no clue which one is best.But I would assume that with spending as much money as they did on the survey, they did discuss the analysis of the results with someone who actually knew what they were talking about.&nbsp; Not just a bunch of second guessing Monday morning quarterbacks who don't really know the details of their internal analysis.</blockquote>

[QUOTE="Oofta, post: 9407909, member: 6801845"] With a scale of only 1 thru 5, I just don't see the 1s or 5s having an outsized impact. It's not like we're looking at a dozen values that between 1 and 1,000. Taking the median when you have millions of responses does not mean you ignore all the results between 2 and 4 any more than if you had used the average. I also haven't seen a quote that said they used the median, although I could have missed it. Sometimes you want to use the median, sometimes you want to use the average for analysis, neither is preferable. Since I'm not a statistician I have no clue which one is best. But I would assume that with spending as much money as they did on the survey, they did discuss the analysis of the results with someone who actually knew what they were talking about. Not just a bunch of second guessing Monday morning quarterbacks who don't really know the details of their internal analysis. [/QUOTE]

Verification