In order to value non-market goods, economists estimate individuals' willingness to pay (WTP) for these goods using revealed or stated preference methods. We compare these conventional approaches with subjective well-being (SWB), which is based on individuals' ratings of their happiness or life satisfaction rather than on their preferences. In the context of a quasi- experiment in urban regeneration, we find that monetary estimates from SWB data are significantly higher than from revealed and stated preference data. Stigma in revealed preferences, mental accounting in stated preferences and unspecified duration in SWB ratings might explain some of the difference between the valuation methods.