Accurate prediction of geopolitical outcomes is vital for evidence-based decision making. Crowdsourcing human forecasts and machine learning models each increase accuracy and reliability of predictions, however, neither method is consistently superior to the other. Hybrid forecasting – pairing human forecasters with machine models – improves predictions by capitalizing on the unique advantages of each method. Our results show that the strengths of machine models and crowdsourced predictions can be simultaneously leveraged to improve accuracy, but distrust and confirmation bias impede performance (Quantifying machine influence over human forecasters). The objective of this research is to compare prediction by statistical models and humans under different conditions (e.g. if contextual framing affects accuracy) and develop better hybrid forecasting systems.