This research investigates how the public of a middle-income country, Thailand, values ecosystem services associated with irrigated rice agriculture using a choice experiment. The results show a significant willingness to pay for services such as drought mitigation, water quality and the environment and maintenance of rural lifestyles and rice landscapes. The iterative procedure developed to fully analyze the incidence of attribute nonattendance (ANA) improved the model fit when compared with a multinomial logit model or an ANA model with potentially only one attribute ignored at a time (ANA-1). Moreover, the inferred probability of the class of respondents having attended all attributes was 45%, compared to 9% with ANA-1 model. However, it also suggests that 55% of the respondents made their choices by considering only two of the five attributes. Finally, this research also suggests that failing to consider ANA does not change the public ranking of scenarios contrasted by the services they would provide but would overestimate the WTP for these scenarios.