In Chapter 2, we make the point that that empirical researchers rarely directly test their favored theories; they instead assess whether their data are consistent with the observable implications (hypotheses) following from their theory or with rival explanations.

Yet even assessing these manifestations of theory requires hard work. The problem is that observable implications, along with rival explanations, are conceptual claims about the relationship between or among variables. Suppose we think that an independent judiciary will lead to greater economic freedom. That's a fine hypothesis but to test it we must move from abstract ideas ("independent judiciary," "economic prosperity") to something far more concrete.

This movement from the abstract to the concrete is known as measurement or, more formally, the task of comparing an object of study (for example, a real world event, subject, or process) with some standard. We could measure a country's economic freedom by asking financial experts to tell us whether they think the country provides economic freedom; or we could examine the number of steps that a start-up business must take to obtain legal status. Either is a measure of the concept of economic freedom, a concept that we cannot observe without clarifying it.

In this chapter, we consider, first, the task of measurement and, second, the standards empirical researchers (and their readers) use to assess the measures they've developed to proxy the variables in their hypothesis. To provide an example: How should we decide whether asking experts or examining the number of steps is a "better" measure of economic freedom?

- end of excerpt -