I presume I need to change the way I think. What I'm actually sampling is error and no error. There are only two possible outcomes. Assignments with or without errors over period of time. My sample size last year was 581 assignments. Of those there were 72 ( 12%) assignments with errors. Can I, presumed random assignments, calculate the confidence level for 12% error in the assignments, given 5% tolerance level, based on my sample size n=581. In other words, will my sample size be enough to say anything about the overall number?