198 ACCESSING AND GENERATING DATA
THE CONCEPT OF PROTEST AND THE
INTERPRETIVE ALTERNATIVEIn early efforts to numerate political protest events, such as that for the 1972 edition of the World
Handbook of Political and Social Indicators, the authors showed an awareness of many of these
problems (C.L. Taylor and Hudson 1972, 62–67). Nevertheless, their numbers are widely used
today as “brute data” without reference to the qualms the coders may have had about the simpli-
fications they were making. C. Taylor (1979, 40) refers to “brute data” as “actions which can be
identified beyond fear of interpretive dispute.” It is quite clear from the many examples of infor-
mation “left out” that the same problem affects Banks’s data. Information on the instigators,
participants, objectives, legal environment, magnitude, impact, territorial distribution, and many
other aspects are not reflected in the numbers. Furthermore, the categories used fail to encompass
major forms of political protest used in India. Indeed, to identify what Banks calls riots, strikes,
and demonstrations as “brute data” is inaccurate—or meaningless.
Part of the reason for the simplification of political protest to numbers is the ease with which
they may be related to other numerical representations of phenomena to develop generalizations.
In positivist social science, developing generalizations is the goal. They are what gold is to the
prospector. Theory is simply a set of interrelated and confirmed generalizations; explanation
requires generalizations in the form of covering laws, and prediction requires them as well. Yet,
the illustrations above suggest that the simplification is a mistake: Banks’s numeration of con-
cepts of protest equates very different events and deprives the observer of information essential
for the building of knowledge. Indeed, when a protest event in India is reduced to a “1,” it is made
virtually meaningless.
Those who defend the utility of data sets argue that the missing data are not intrinsic to the
concepts. Rather, they are pieces of information that may be treated as external to, or separate
from, these concepts. For example, the success of a demonstration may be numerated apart from
the occurrence of a demonstration, or the objective of a strike may be numerated apart from the
occurrence of a strike. Since the “success” or the “objectives” may be detached from the occur-
rence of a protest event, their absence does not detract from the empirical value of the concept. In
other words, the “missing” information may be treated as variables whose relationship with pro-
test may be empirically tested—or it may be treated as residual variables assumed to have no
impact on protest.
Furthermore, much more sophisticated protest data sets have been developed that include the
“missing information,” that is, the criticisms of the Banks data are not generally applicable. An
example would be the protest event data collected on Eastern Europe by Ronald Francisco.^61 He
has gathered a wide variety of information on each event—much of which, as I have argued, is
missing from the Cross-National Time-Series Data Archive. He collected information on where
the protest events took place, when they took place, what the issue was, how many protesters
were present, the damage done, the size of the state forces dispatched to deal with them, and many
other dimensions. Thus, much of the “missing information” might be obtained.
The core contention of both challenges is that the information cited here as “missing” is not a
part of the meaning of the protest concepts, that is, that its absence poses no problem for those
who wish to understand the nature and impact of political protest in India. Such a contention is
contradicted by the data presented here. First, the categories exclude important protest actions:
Major forms of expression of political protest, such as individual fasts or suicides, are excluded.
Second, the categories do not correspond unambiguously with major forms of political protest,
such as the bandh or a jail bharo. Third, and most importantly, to treat all these pieces of informa-