Mitigating Bias in Venture Capital

Based on review from literatures, 11 bias plus 6 sub-bias, which are similarity bias (including religion background, syndication/network ties and social/personal ties), continuation bias (including sunk cost effects, escalation of commitment and status-quo bias), self-attribution bias, overconfidence bias, reputation bias, risk propensity bias, visual cues bias, linguistic presentation bias, mood/emotion bias, look-ahead bias and local bias, during decision making process from venture capitalists were summarized in this thesis, impacts on decision making process were indicated, available measures against specific bias were also mentioned. Furthermore, four traditional methods, which are Bayesian casual maps, repertory grid, analytical hierarchical process and analytical network process, to mitigate the overall bias were illustrated. Going forwards, some computer-based methods with help from machine learning were then being pointed out and described. At last, a conclusion was made, limitations and future works were outlooked.