Does choice-matching make a difference? Incentivizing stated preference responses in the field