ABSTRACT
It is clear that if we take a sample of patients from the population of interest, then the true population mean (µ) cannot be determined. However, we may estimate the population mean by taking the mean of the sample (called the sample mean, denoted by ). The bigger the sample used, the closer the estimate is to the population mean (see ‘Sample size and power’, p. 50). In clinical practice it is rarely necessary to know the precise population mean. For example, knowing that the sensitivity of a diagnostic test was 92. 13% yields the same clinical decision as knowing that the sensitivity of the test was between 91% and 93%. In addition to the population mean, it is also important to consider the variability between observations in the population. As discussed in ‘Measures of central tendency and dispersion’ (p. 83), the population standard deviation (σ) provides a useful measure of variability. It can be estimated using the standard deviation of the sample (called the sample standard deviation, denoted by s).
