Interpretation and Method Empirical Research Methods and the Interpretive Turn

188 ACCESSING AND GENERATING DATA

counts right, the numbers would be meaningless. So much that was different among the events

had to be ignored in order to categorize them as riots, strikes, or demonstrations that these con-

cepts no longer meaningfully reflected an empirical reality. The chapter that follows describes

what is left out when protest events are numerated and suggests that the way forward may be

through an “interpretive turn.”

❖❖❖

It appeared that the word “strike” at times meant something different than a strike as

defined by Western scholars. And a vast array of other terms were commonly used to

refer to different types of protest action. To put all of these protest actions into “riot,”

“strike,” or “demonstration” deprived the actions of their meaning. That meaning is not

only culturally determined, but situationally determined, too.

—Dean McHenry, personal reflection

THE PROBLEM

A number representing a summary of the frequency of forms of domestic protest for a year in a

country of over a billion people, like India, must substantially simplify reality. The advantage of

numeration is that it allows the application of highly sophisticated statistical techniques for pro-

cessing information. The disadvantage lies in the loss of information. In this chapter, I will argue

that the numeration of riots, strikes, and demonstrations leads to such a significant loss of mean-

ing that the use of these broad categories is unlikely to lead to the development of knowledge.

Such a numeration may, in fact, misinform and mislead.

The challenge is to find an alternative methodology that will facilitate the creation of knowl-

edge. A logical alternative is an interpretive approach. It builds not from phenomena simplified to

virtual meaninglessness, but from phenomena portrayed in their complexities. It uses reason and

logic both to provide for missing information and to draw more informed categories from the

complex body of information assembled.

I will examine this issue by focusing upon the conceptualization used in the quantitative index

of domestic protest in India developed by Arthur S. Banks and placed in the Cross-National

Time-Series Data Archive (CNTS).^1 The CNTS contains a vast array of data sets covering most

countries of the world over a period of many years developed from a variety of sources. My

concern is with three of his measures of domestic conflict: general strikes, riots, and antigovern-

ment demonstrations. The source of the numbers Banks reports is the New York Times. The justi-

fication for the selection of this index is that it is widely cited in the political science literature,

that it embodies the basic characteristics of other numerical indices, and that it deals with a matter

of particular importance to India. Furthermore, the size and complexity of India is likely to make

more obvious the problems associated with numerating and interpreting the three concepts than

would a smaller or more homogeneous country. In the concluding section an assessment will be

made of whether interpretivist approaches “put back in” what numeration “leaves out.”

THE CONCEPT OF DOMESTIC PROTEST AND ITS

NUMERICAL CHARACTERIZATION

Banks’s CNTS includes a simple compilation of the frequencies of demonstrations, strikes, and

riots. Although scholars focusing upon events analysis in other parts of the world have sought to