Compute the mean, or average, of the data points. This is achieved by adding up the all values of the data points and then dividing the sum of those numbers by the number of data points. If there are five data points, add the five data point values together, and then divide by five.
Find the variance. Take the difference between each data point and the mean, then square that difference. If the mean is four, and the first data point has a value of six, the difference is two and the value would be squared producing four. This is done for each data point, the results are added together, and then the total is divided by how many data points there are.
Compute the standard deviation. Take the square root of the variance, producing a smaller number known as the standard deviation.
Find which values are normal. Add and subtract the standard deviation from the mean. This provides a "band" in which approximately 68 percent of the data points will fall if the data are normally distributed. By multiplying the standard deviation by two, the value for two standard deviations will be produced. Adding and subtracting this value from the mean will provide a band in which approximately 95 percent of the data values will fall. Within three standard deviations of the mean approximately 99.7 percent of the data values will fall, assuming a normal distribution.
Isolate outliers. Find the data points which are outside of two or three standard deviations from the mean. By bringing these points toward the mean, the variance and standard deviations can be lowered. Moving all or some data points closer to the mean will also reduce variance and standard deviation.