the Poisson probability distribution, then the data follow a Poisson
probability distribution (Gupta, 1977). Here, based on the occurrences
data-set and the frequency information in Table 3 and Fig. 2,
the Pearson’s goodness-of-fit Chi-square test statistic can lead to
acceptance of the Poisson model at the 0.05 level of significance.
Under the Poisson probability model for the number of occurrences,
the mean and variance must be equal, and the mean can be
expressed by a linear equation; if the variance is larger than the
mean, the sample data is overdispersed, and if smaller it is
underdispersed. Under such circumstances, adjustments must be
made using a rescale in advance so that more accurate standard
errors of the estimated Poisson regression parameters and subsequent
p-values are obtained. Here, the estimated mean of m is the
average number of occurrences, y ¼ 2:037and the estimated
variance is 2.019 (Table 3), and the ratio of the variance over themean is 0.991, which may indicate some underdispersion. To test
for this we consider whether the Poisson variance is smaller than
its mean using Equation (1) to check whether a ¼ 0 where g (m) ¼ m2
(Cameron and Trivedi, 1998).