Confidence and measure of speed

It is important not only to know measures of location (the mean), but to get some indication of the spread of data. Measures of spread (such as the standard deviation, standard error and variance) lead us to be able to estimate errors in measured quantities. For example, if we measured the sensitivity of a set of microphones for a manufacturer and found them to give very different values, this would indicate a great variation in the build quality. Crucially, we could not be very confident about a single microphones sensitivity if only the mean value was given. We need information about the spread of data to get a handle of the likely range of values that this type of microphone provides.

Find out more about measures of spread and confidence limits (it is in PDF because it incorporates equations which don’t work well in html).