Risk assessment based on probability-impact (P-I) ratings is the most widely used approach in project-based industries such as the construction industry. However, there are various criticisms about utilization of the P-I rating approach because there are factors and assumptions that are hidden within the risk ratings that cannot be conveyed to decision makers for a reliable risk assessment. This study aimed to explore biases, particularly how the risk attitude and assumptions on controllability of risk (illusion of control bias) may affect subjective risk ratings assigned by the experts during risk assessment of international construction projects. Results showed that as the level of perceived controllability rose, risk ratings assigned by the experts tended to be lower. There was a moderate correlation between risk attitude and risk ratings. Risk attitude and assumptions on controllability were also moderately related, and their combined effect on risk ratings varied according to different risk scenarios. Risk ratings were affected by the risk attitudes of experts, especially when the country risk level was high, whereas assumptions on controllability tended to affect risk ratings more significantly when the country risk level was low. Although the questionnaire findings about the impact of biases on risk ratings are valid only within the context of this study, findings may have some generic implications for developing new methods that can highlight and control the hidden factors in subjective risk ratings. Assumption-based thinking and knowledge elicitation on risk ratings, together with underlying assumptions by group decision making, may decrease the impact of illusion of control bias during the risk assessment process.