A good way to produce a more valid survey (reduced "ballot stuffing and sample bias) would be if Daz (Rendo, etc) asked everyone to fill out something similar immediately following the completion of their first purchase after the survey was introduced.

Of course, we'd be unlikely to see the results...

The way questions are posed will affect results dramatically, as will the order in which they are asked. Respondents can be "primed" as well, by the environment in which the survey takes place. It's hard, if not impossible, to create a survey that doesn't introduce some bias, even in if unintentional.

I didn't mind this survey, even though sample bias is always going to be a problem when the sample is completely non-random

It would have been interesting, though, if a question had asked how much each person spent on average at Daz. If 90% of the $10 a year people stopped buying, it might not affect sales quite so much if 90% $10,000 a year people continued.

 

 

---

PoserPro 2014, PS CS5.5 Ext, Nikon D300. Win 8, i7-4770 @ 3.4 GHz, AMD Radeon 8570, 12 GB RAM.