Suppose that 2,000 scientists compute 90% confidence

intervals for parameters they are interested in. How many of
these intervals would we expect to successfully capture their
parameters