Lecture14570统计笔记.docxVIP

PAGE  PAGE 13 Confidence Interval for one population mean or one population proportion, continued 1. Sample size estimation based on the large sample C.I. for p From the interval are given and we are interested in sample size . Therefore, (When , it has the maximum value.) Example 1. Example 2. 2. Sample size calculation for p based on the maximum error E. Definition. We want to estimate within with a probability of . Derive the formula for n Pp-p≤E=1-α P-E≤p-p≤E=1-α P-Ep1-pn≤p-pp1-pn≤Ep1-pn=1-α Since when the sample size is large, by the CLT we have: Z=p-pp1-pn~N(0,1) Thus P-Ep1-pn≤Z≤Ep1-pn=1-α c=Zα/2=Ep1-pn ∴n=Zα/22p1-pE2 When p is unknown, we can plug in the estimate of p, p, and obtain the following formula: n=Zα/22p1-pE2 Recall we also derived based on - the length of the % large sample confidence interval for . Their relationship (using the second formula for sample size calculation because our CI formula was basd on the second PQ where we estimated the p in the denominator) is Recall we also derived based on - the length of the % large sample confidence interval for . Their relationship is The % confidence interval for is . The length of the confidence interval is Example 3. In order to estimate the percent of children with inadequate immunization to be within 0.05 of the true proportion with a probability of 98% How many children should be sampled? Solution. If the percentage of children with inadequate immunization is estimated to be , then Solution. Sample size calculates for 1 population proportions based on the maximum error E. 3. Confidence Interval for 1 population mean There are 4 scenarios – we cover only the first 3 scenarios in our class. Normal population, is known. Point Estimator : Pivotal Quantity : CI for : Length of CI : Sample size based on : Sample size based on Normal population, is unknown. Point Estimator : Pivotal Quantity : CI for : Length of CI : Sample size based on : Samp