Suppose that the exact values of the data x1, ..., x50 from an exponential distribution EXP(theta) are unknown, but it is known that 40 of the 50 measurements are larger than t.
a) Find an approximate one-sided lower 95% confidence limit for P(X>t) based on this information.
b) Note that under the exponential assumption, P(X>t) = exp(-t/theta). If t=5, use the result from (a) to find an approximate one-sided lower 95% confidence limit for theta.