
06617841
j
2016e.00914
Farnsworth, David L.
Impact of misclassification on estimation of a single proportion.
Math. Comput. Educ. 50, No. 2, 109113 (2016).
2016
MATYC Journal, Old Bethpage, NY
EN
K70
measurement errors
stochastics
probability theory
conditional probability
random variable
binomial distribution
misclassification probabilities
error probabilities
confidence interval
estimation theory
statistical inference
point estimates
From the text: Measurement errors are a constant problem to experimenters yet they are little mentioned in elementary statistics courses. On a topic separate from Bayes Rule, misclassification rates can and should enter our calculation when estimating a proportion. The formulation is in a binomial setting, and we use the largesample normaldistribution approximation. We present a point estimator and a confidence interval for a single proportion when there might be misclassifications. The analysis is surprisingly simple and can be introduced at any level statistics course. The ideas can be used in the classroom or in easily constructed exercises.