I'm probably coming in late on this one (since the paper first came out in 2009!), but I recently read this paper from Phil Graves. I got the chance to ask a few questions about it online and was intrigued by the consequences. I'm curious to see what other people who are regularly estimating WTP/WTA think about these consequences - namely that you're estimating at the wrong income level for some of your survey respondents (likely a greater number of your respondents than representative given response rates for "those who care"). Has anyone tried the idea that Graves mentions in the LinkedIn questions where you scale your response to the private goods-equivalent gap?