ABSTRACT
The most common form of industrial sampling is the case in which momentary observations are made at randomly determined times. This methodology is typically referred to as “work sampling” or “occurrence sampling.” The basis of this methodology is simply binomial or Poisson sampling, which are well described in many books on statistics. In its simplest form, an event either occurs at an observation or not, and if that event occurs on 40% of those observations one would expect the event to occur with that frequency plus or minus sampling errors. Forty percent of 10 observations clearly has a much lower statistical precision than 40% of a 1000 observations. That precision is determined by the statistical confidence interval from the sampling.
