15 and n(1 – p) Which is half the width of the confidence interval. How Do We Choose the Sample Size for a Study? Suppose we want to construct a 95% confidence interval on p, where the margin of error, m, is selected prior to drawing the sample. T-table, p. A3 Confidence Level 80% df . . . 50 60 80 100 What sample size, n, is needed? Solve for n in m = 1.96 p(1 – p)/n ^ ^ to obtain n = p(1 – p)(1.96/m)2. How Do We Choose the Sample Size for a Study?