ABSTRACT
The procedure for interval estimation in discrete distributions is the same as that for continuum-valued variables, except that since CDF and X-values can only have particular values, the CDF and/or x-limits need to be truncated (up or down) to the best nearest feasible value. It could very well happen that the people have a bunch of data (perhaps more than 1,000) representing a process, do not have a model for the distribution, and are asked to determine the upper and lower limits for either the data of the range of the distribution parameter values. The question in this chapter is not what might be the data range that the population could generate, but what range of population parameter values might have generated the dataset. Traditionally, the default is the 95% confidence interval with equal allocation of the extremes.
