We can never know with certainty the true value of an intervention effect because of random error. It is inherent in all measurement. The observations that are made in a study are only a sample of all possible observations that could be made from the population of relevant patients. Thus, the average value of any sample of observations is subject to some variation from the true value for that entire population. When the level of random error associated with a measurement is high, the measurement is less precise and we are less certain about the value of that measurement.