We used both stated preference and revealed preference data to estimate willingness to pay (WTP) for shade‐grown coffee as compared with conventionally grown coffee. Stated preference data was collected using contingent valuation studies. Revealed preference data came from an experiment where all survey participants received a personally identifiable voucher redeemable for a free bakery item when the holder purchased a coffee. We compared estimates of mean and median WTP a price premium for shade‐grown coffee from stated preference data with similar estimates from revealed preference data. We used a logit model to evaluate the effect of explanatory variables (measures of environmental attitudes, personal norms for pro‐environmental behavior, and demographic variables) on respondents' WTP a price premium for shade grown coffee. Model parameters were estimated using the maximum likelihood approach. Respondents with higher scores on measures of environmental attitudes and personal norms for pro‐environmental behavior were, on average, willing to pay more for shade‐grown coffee. While this paper examined a specific case, purchase of shade‐grown coffee, our results confirmed that stated environmental concern was a good predictor of pro‐environmental behavior. We found that mean and median WTP estimates from stated preference methods were higher, but not significantly different than mean and median WTP estimates from actual purchases, indicating convergent validity between stated and realized preference methods. The majority of individuals both stated WTP a price premium and purchased shade‐grown coffee at a price premium. We did, however notice some interesting behavior at the individual level where stated preferences under‐predicted realized preferences at low price‐premia and over‐predicted realized preferences at high price premia. This article is protected by copyright. All rights reserved.