Series of measurements are taken to make sure that the experiment is able to measure different sets of conditions under which the person or the instrument has to work.
Now the question arises how to make sure that the series of measurements are not only reliable but valid as well. Or can we make use of a measurement which happens to be reliable but low on reliability or vice-versa. This study cannot be an area specific but it cannot be generalized for all experiments as well. There might be some areas where we can use such measurements, but there might be areas where we need to be doubly sure that the measurement is both reliable as well valid. It is worthwhile here to mention that while validity can be tested and can be termed as an objective parameter, it is difficult to do exact calculation for reliability as it is subjective in nature. Reliability is therefore estimated in order to come to the nearest perfection point. Four general classes of reliability are1;
While there's no doubt that a measurement with high validity and high reliability is the most sought after type of measurement, a measurement with low validity and low reliability is never used for making any calculations. Measurements low on validity and high on reliability imply that we are consistently measuring wrong values, which in a way defeats the very purpose of carrying out the research. In such cases we continue to get incorrect information quite consistently. Therefore, we seldom use such measurements. Quite often it so happens that the repeated measurements fail to show consistent patterns, but for a group of parameters, the measurements hold good. In this case, though on an average we do get the valid results, the consistency is not there. In such cases, the measurements can be useful depending upon the criticality of the usage. For example, if we are to make use of these measurements for the operation of