Experimental quantity, mental budgeting and food choice: a discrete choice experiment application

Abstract Food discrete choice experiments typically define product alternatives with a researcher-predetermined and sometimes arbitrary quantity. Results reveal that the use of a researcher-prespecified experimental quantity leads to biased welfare estimates. Differences in marginal utility of money are found with a resulting upward bias in willingness to pay estimates when small pre-defined product quantities are used. Higher-income consumers show more evident bias. This evidence cautions the use of a researcher-predetermined quantity to design alternatives in choice tasks and also proposes an alternative experimental design that accounts for these effects by matching the quantity in experiments to consumer’s actual purchase quantity.