How to make your respondents' answers more realistic?

Hello all,
This is my first post in this forum, so I am a bit excited...
I am a Ph.D. student in Transportation Planning and I use marketing models to analyze choice behavior. My research tool ia a Qualtrics questionnaire which I designed and based on the data from the questionnaire some models will be used.

I have a methodological problem which I am not sure how to solve - so I hope someone here could give me a new insight.

In my research I am interested in choice behavior under uncertainty. How people act when they do not know how the product details will change over time. My application is airline itineraries, which their price and availability is changed according to airline's yield management. In my research a respondent is provided with a certain time period with daily intervals. In each "day" he/she gets a different choice set. So a respondent starting a 30 days period receives the first set which contains 6 itineraries of different airlines. He/she can do the following:
1. Choose an itinerary - that would end participation; or
2. delay decision to the "next day" - gets a new set that might be completely different.
A respondent is free to delay his/her decision as many times as he wants until his departure day. So the maximum sets presented to a respondent is up to 30.

Now here is my problem. In real life, people choices has implications - both monetary and usability. A person wants to fly so he is trying to get the best product he can find. Furthermore, money is spent and purchase cancellation is not always possible. When filling a questionnaire similar to what I have, the utility is different. Nothing is really purchased, a flight will not occur, and money is not a real issue. So the utility might be to finish the questionnaire as fast as possible. So a person which in real life will spend a week to find the best deal might not do so in my questionnaire (in this case postpone his/he decision 7 times). That will create a seriouse bias.

I am trying to find a way to manipulate my respondents so they would act more realistically and minimize the fatigue. Up till now I am not sure that I have a good solution. I am telling the respondents that their chances to win the raffle - we have cool prizes, like Android Tablet computers and gift certificates - are linked to the quality of their choices (we have a way to determine that). The problem is that it might not work the right way. I thought on setting a fake budget for each respondent but I am afraid that it might narrow the choice to cost-based only. So the budget sollution is not implemented.

Beside that I am stuck. Can anyone please help me here to solve this problem?

I am gratefull for any help,

Uzi