This test is also known as the binomial distribution test. The test is used to consider the proportion of a sample for which a particular qualitative observation has been made. For example, the proportion of rolls of a die that have come up 6. The test examines how far from the expected proportion is the observed proportion, given an assumed probability for the observation. In its use in this paper, we adopt a null hypothesis that two planners should win with equal likelihood and test the proportion of observed wins for one planner against this hypothesis. If the deviation of observed proportion from expected proportion is sufficiently high, then the null hypothesis can be rejected.

Derek Long 2003-11-06