Freddies Voting Very Confusing
 

Very counterintuitive and contradictory, that you'd be asked to select the "best" airline/hotel program, and then be asked to rate it from "best" to "worst."

If you're selecting a carrier/property to be best, why provide the option for "worst?"

What if you think all of them stink, but this is the best of the "bad" bunch.

In that case, you'd chose that one, and rate is as worst.

I completely and wholeheartedly agree with you. I am a market research professional and spend the better part of my week creating surveys that will be given to consumers. It seemed clear to me that whoever designed the Freddies survey really had no concept of how to design an accurate and effective survey.

I own a marketing research firm and tend to agree that a straightforward likard scale on all available programs and properties would have been less confusing. That way, you could have easily ranked any program you were familiar with and had an implied or selected 'no opinion' on the rest. By asking us to pick the 'best' and then offering a 1-10 likard scale, it's not clear what conclusions you can draw from the subsequent analysis. For instance, if I pick USAIR as the 'best' and then rank it a 1 on the likard scale, does it now rank lower than a program I didn't choose as the 'best'?

In spite of the opening instructions "Which individual frequent traveler program offers the best overall combination of service, bonuses, awards, benefits and opportunities?", Are you really voting for 'best', or just the program you want to rank? I say this because of the plans of the "deny continental the freddie" movement. By selecting CO and ranking it a 1, have you in fact ranked it low, not high? As long as your first click as 'best' has no statistical consideration, that's one thing. But if the first click does have statistical consideration, I think the results are bound to be confusing. And if I'm not sure how a ranking I submit is being addressed, I tend to think that confusion is pretty high, which in itself will affect results.

I commend Randy for conducting this and I'm aware of how thin he is spread in terms of personal attention. I don't intend to be overly-critical, but IMHO, I would agree this could have been designed in a manner that would produce statistically conclusive results.

[This message has been edited by svpii (edited 01-28-2001).]