This paper compares willingness-to-pay (WTP) estimates elicited using three separate methods: in-store experimental auctions, a mailed survey with a cheap talk script included, and the same survey with no script. The products in question were four steaks bearing hypothetical brands representing various brandable attributes. It is found that WTP elicited using experimental auctions was the lowest, followed by WTP elicited by the mail survey with a cheap talk script, then the mail survey with no cheap talk script. Tobit and double-hurdle econometric models are used to identify factors influencing respondents’ WTP; model results are largely consistent with previous findings.