School? Degree?

University of Cincinnati in the MBA program at the moment with a concentration in Marketing. Once that is complete I'll go for my masters in Business Analytics.

One challenge with writing (and answering!) survey questions is that they often embed assumptions which, depending on how they're interpreted, can significantly shift the answers a person gives.

Questions 11-12 ask how much more I'd pay to for a ticket on my favored airline as opposed to a similar routing on another airline. I put large dollar amounts in as my answers because while the notional benefit of "loyalty" is worth very little to me financially, the quite tangible benefits of Platinum Premier status on UA are free access to Economy Plus seating and a spot on the upgrade list to First Class. On airline "B" I receive no such benefits so I'm willing to spend more for UA.

Questions 13-14 ask how much more I'd spend for mileage runs. Again, I put in fairly large dollar amounts. Here the reason is that the amount I'd spend for an MR on airline "B" is ZERO! An MR is almost never worth it outside of pursuing status on a preferred airline.

Even Question 30 bakes in an assumption. "You're on a plane with 7 rows in First Class. Seats AB on the left, CD on the right. Which seat would you select?" Answer: really depends on whether the primary cabin door is forward or aft of the seats.

It's funny you mentioned that, that's part of the reason I put 30 questions in the survey over multiple pages. I'm really only trying to correlate a very small percentage of the questions, so I had to try to trick the respondent. Some of the ancillary questions are there to help find outliers and/or identify people who are trying to "cheat their answers" so their responses can be disqualified.

Unfortunately this class (essentially spreadsheet modeling) is coming a semester before my market research classes which get in to exactly what you were talking about.

I'll message you privately about one of the points you made because you just pinpointed something we were cracking up laughing over.

Completed, one issue is that the mileage run questions are labeled optional but it forces an answer, so mine were meaningless (dont do mileage runs)

Completed. But I'm another respondent who doesn't do mileage runs. No option for n/a there.

Really? Let me go back and double check. I'm getting the appropriate number of skips on the MR questions but maybe one was tagged as mandatory.

Edit: Got it, I see those now.

I started to help you out, but I am not willing to disclose income information, and your survey would not let me skip the question or choose "prefer not to answer"

If you fix that, I'll go back and finish the survey for you.

Can not unfortunately as it will result in inconsistent data since the previous takers had to make a selection.

All questions should have ability to choose 0 or none of the above.

Completed. I'm also quite interested in the results!

As someone in the research world, I'm just going to throw out there that skipping a question and answering "none" are usually two different participant responses. This will probably come up when you take that market research class next term.

Q 13 is still forcing a response but did accept the non-numeric "-" so a heads up to recode the string to a missing value.

Also, Q 17 & 18, you're missing a word: "...in order to fly one of the airlines..."

And with your income question I highly recommend in future surveys that you specify whether you want individual or household.

HA! God I hope so. The point of this class is really more about processing the data and modeling, so for this project he was looking for something relatively simple. However, since I think big, I decided to make it a little more complex.

Maybe after I complete the degree I'll do the survey again, use it to get a job at one of the airlines and ACTUALLY enhance their programs for the rest of us.

You should add skip patterns in future surveys. In this one, you ask the respondent whether they've ever been on a mileage run at the the. You should move that question to the beginning, then ask detailed questions about mileage runs only only if respondent has been on a mileage run.

Polldaddy is really robust in that regard, but I didn't exploit that functionality. It allows branching but that goes well beyond the scope of what I'm looking for.

As I mentioned before a lot of the questions were added in order to mask what we were looking for, but oddly enough these filler questions are where the really interesting trends are forming. I promised myself I'd stop looking at the results until next Monday the 25th, but I will keep poking in here to bump the thread.

Short and painless.