Paired-Sample Tests Based on t and W

Abstract

Sometimes it makes sense to measure each of the items or individuals forming a sample not once, as has been assumed in previous chapters, but twice, first under one set of conditions, then under another. For example, the gradient of a beach might be measured at the same randomly selected locations both before and after a storm to see if there was any change. Or a random sample of commuters might be approached and asked how long they take to travel to work using (a) public transport, (b) their own cars — the objective being to determine which is the quicker method of travel. Or two different instruments for measuring soil moisture might be tried out side by side at a number of sites selected at random in order to see if there is any consistent difference in the readings.