The path towards an integrative negotiation depends crucially on knowing whether and how the negotiating parties differ either on risk perception or perceived risk attitudes [Weber and Hsee 1998]. An understanding of risk preferences of an individual could prove valuable in the negotiation process and outcome. It would also help the negotiators to better understand their opponents and their issues. Although knowledge of the dynamics of risk taking is still limited, several authors have shown that decision makers can be simultaneously risk seeking and risk averse in different domains and situations, implying that risk preferences are context specific [Slovic 1972; Hershey et al 1982; Payne et al 1980; MacCrimmon and Wehrung 1984; MacCrimmon and Wehrung 1990; Schoemaker 1990; Shapira 1997]. There is little correlation from one setting to another in a persons risk preferences [Slovic 1972]. Only those tasks highly similar in structure and involving the same sorts of payoffs have shown any generality and as similarity decreases these cross-task consistencies rapidly decline. Thus, an individual who takes risks by guessing often on a mathematical exam (under negative marking) is likely to be high-risk taker in other exams as well but that does not imply that he would prefer a high-risk occupation. The context specificity not only relates to the substantive domains (e.g. negotiation outcome versus health outcome), but also to measurement procedure (e.g. response modes or question framing) [Hershey et al 1882; MacCrimmon and Wehrung 1990; Pennings and Smidts 2000]. No single measure of risk preferences is adequate to capture the complexities of risk taking behavior and different types of measures reveal different risk preferences [MacCrimmon and Wehrung 1990]. The objective of the present study is to examine the individual risk preferences in online negotiations. Specifically, to examine whether risk preferences are related across payoff domains (gain situation, mixed outcome and loss situation) and for different measurement approaches and whether these risk preferences share a common variance that can be attributed to a global construct.