Science begins with measurement. For example, to determine the effectiveness of a weed-killer, we must be able to quantify the abundance of the targeted weed in the control and the experimental sample. It is almost superfluous to add that without reliable measurement, the analysis will produce bad science.
Similarly, investment analysts can produce useful findings only if they have valid underlying data. In the real world, however, analysts have to face the hard truth that the data may be unreli