Unfortunately we do not have any stat gurus around. I get to figure everything out today.
No validation of scales per se, however it's basically a mirrored grading scale with Excellent being an A and Poor being an F. Could even break it down to GPA if it comes to that
For the most part, all the results have been consistent with each other. Just need to figure out how these results will help us continually improve customer satisfaction. I want to be able to show management some graphical data rather than hand them a stack of surveys.
I just get concerned with doing a lot with survey information until you have assessed some level of validation of the scores. Here's an example:
You want to measure how the customer viewed the packaging.
You can ask:
The packaging did not adequately served the need (A-F)
The packaging did adequately serve the purpose (A-F)
The product was properly packaged (A-F)
You did not like the packaging (A-F)
You loved the packaging (A-F)
I thought you said you used A-F. However, which one are you more likely to mark-an F, or a 1? I would hardly mark anything an F; that just represents absolute failure.
Now the question, what were you really trying to measure? Answer-to measure if the product arrived safely and intact in the package, with no visible damage on the outside.
However, the four questions seemed to be about packaging, yet they measured slightly different results. People perceive things differently. It's not a right or wrong thing; just different.
That is what I mean when I speak of validating the survey. Before presenting and making a big deal of survey results, you want to make sure you measured what you purported to measure.