ABSTRACT
One can place a thermometer into a beaker of water, withdraw it, read the temperature and then place the thermometer into another beaker. Two temperatures can be compared and we know if beaker A is warmer or cooler than beaker B. If we are told that beaker A has a temperature below zero degree centigrade, then we may be relatively certain that the water was in its solid phase (ice) at the time of measurement. Assuming, that is, that the water was relatively pure, that it was at sea level, and that the thermometer was reliable and valid. A test can be used to compare entities on amounts of some attribute, to assign a value to the amount of the attribute the entity displays that has meaning in some reference system, and to make probability statements that the entity is in some category that has distinct properties and behavior. The example of the temperature of water is useful to remind psychologists that these three uses of measurement are general uses and not limited to the interpretation of scores on psychological tests. In the water example, the prediction that the water was in solid phase depends on two conditions. First, the person making the prediction must have a number of facts in hand that are all relevant to a veridical theory of the phases of water. And second, he must have a measuring instrument that is both reliable and valid for the purpose of measuring water temperature. In this trivial example of predicting whether water is solid or liquid from its temperature and pressure and purity, the observer can predict the state of the water with a high degree of confidence because he has a strong theory of the phases of water and a powerful measuring device in the thermometer. Neither the theory nor the instrumentation suffers from large imprecisions.
