One of the problems with a study such as this one is the small size of the samples. If you wanted to

Question

estimate the proportion of a certain color of candy to within 2% of the actual proportion, how many
individual pieces of candy must you sample? Assume you want to be 90% confident in your results and
no previous information is known regarding the proportion values. Based on the total count of candies in
your package, approximately how many packages of candy would be required to obtain this minimum sample size?