It is often the case that two or more economists, observing an identical set of macroeconomic data (national income accounts, inflation, unemployment), arrive at very different explanations of events. How can this be accounted for? Could this occur in a natural science?