If we were making inferences from a single sample, we would expect that there would be a percent probability that the sample mean or proportion generated from our sample results would be within two standard errors of the true population mean