Which of the following would be the best interpretation of a 90% confidence interval? For a given sample, 90% of the observations would be contained in that confidence interval. If we were to (theoretically) take 1000 repeated samples of the same size, then approximately 900 of the CIs would contain the true value of the mean and approximately 100 would not. There is a 90% probability that a particular CI we have computed contains the population mean. If we were to (theoretically) take 1000 repeated samples of the same size, then approximately 950 of the CIs would contain the true value of the mean and approximately 50 would not.