ABSTRACT
The data were collected by Dr A. Barr, Oxford Regional Hospital Board. Interest lies in any systematic variations in arrival rate, especially any that might be relevant in planning future administration. These data represent point events occurring in time. A very simple model for comparison with such data is the completely random series or Poisson process. The simple Poisson process is nevertheless valuable in judging the precision of comparisons, e.g. between different days of the week. The Poisson distribution provides a standard against which the statistical significance of the observed variation can be judged. Significance of variation above that to be expected from a Poisson distribution is tested by dispersion test that is in effect by comparison with fitted frequencies assuming constant rate. The exponential distribution corresponds, however, to a Poisson process of constant rate and the effect of variations in rate is to increase the ordinate of the probability distribution at the origin by a factor approximately.
