There is a 95% probability that the population mean $\mu$ lies within the specific calculated interval.

95% of all sample means lie within this interval.

If we repeat the sampling process many times, 95% of the constructed intervals will contain the true population mean $\mu$.

There is a 95% chance that a single observation from the population falls in this interval.