"I am 95% Condent that u is between 40 and ED" does not mean PHD-sus]: (1.95 because the u is not a random E and does not vary. It is the E that is random because the sample is assumed to be random. To have 100% confidence, the range would have to include possible values and would be totally useless as an estimate.Variability of a sample estimate not only effects how close it is to the parameter, but also the confidence in the of the parameter relative to the estimate.If we have a sample of categorical data, the empirical probabilities in a contingency table are estimates for the probabilities (i.e. population proportions for each group)