Investing in the Stock Market: Statistical pooling of individual preference judgements

Literature concerning the quality of individual and face-to-face group judgements has generally concluded that both groups and statistically pooled individuals outperform randomly chosen or average individuals. This paper extends previous research by comparing statistically pooled individual judgements of both individuals and face-to-face groups in a stock selection task. In general, decisions that would have resulted from statistically pooled judgements were better (as assessed by future stock value) than those that would have resulted from individual or face-to-face group judgements. In choosing among pooling methods, majority rule is often thought to be a very compelling criterion. However, majority rule can produce intransitive group preferences. Methods that use some procedure to resolve the intransitivities of majority rule did not perform well relative to other non-majority rule based methods. Another class of pooling methods, termed equity methods, used in conjunction with ordinal judgements, are recommended based on simplicity, performance, and fairness criteria. The results are discussed in terms of the nature of the task.