Before you float a poll or an exam question you put it through "peer review" to sort the bugs out.
Perhaps, start the question .... "for non-industry pros" .... that way you weed out all the shop techs, instructors and associated types.
The other "first thing" to do, is to see if your little "research universe" is representative of the target population. Are the users of EpicSki representative of skiers in general or "at large"? Probably not! The users of this forum are probably a bit more dedicated and prone to buy more, take more lessons ..... or are in the industry. This is a basic sampling error problem. The only way to mute this problem is to expand the "universe" or change the collection point, perhaps collecting data at the ticket window or when booking reservations.
Destination resort versus feeder area would be another variable.
Most of our customers who are in the Methodist Choir group or the "weekend outa' the city for a romantic interlude" will never be back for a second shot. Populations within populations are a problem if you sample during the wrong period or only in a single period.
Once you conrol for as much "static" as you can, you can also increase the validity by increasing the sample size. The problem here is that you will never get any kind of numbers (at Epic) tht are in any way representative. Heck, one log in from our very own "disski" from Oz would skew the numbers. That woman takes a lesson-a-day whether she needs it or not. She IS a professional lesson taker.
It's a rare person who can ask a simple question and get the data they want with minimal "static". That's why you pass it around in the lunch room. That's us! [img]smile.gif[/img]
Sorry ..... probably more than needed .... but I feel better!
:[ November 25, 2002, 05:07 PM: Message edited by: yuki ]