USING STANDARD DEVIATION, STANDARD ERROR AND RELATIVE STANDARD ERROR
Because of its close links with the mean, the standard deviation can be seriously affected if the mean is a poor measure of central location. The standard deviation is also very sensitive to observations well outside where the majority of observations are found (outliers); therefore, it is most useful in regularly distributed sets of data.
The Relative Standard Error is used as an indication of the reliability of the data. Let us look at the data below extracted from the Survey of Persons with Two Cars.
survey of number of persons with 2 cars
persons with two cars
|persons with 2 cars |
|persons without 2 cars |
In the case above a RSE of 40% for 10 people who have two cars means that the true value could be four persons either side of 10 persons. So the true value is likely to be between six and 14 persons.