THE NUMERATION OF EVENTS 187187CHAPTER 10
THE NUMERATION OF EVENTS
Studying Political Protest in India
DEAN E. MCHENRY, JR.
My graduate training reinforced my sense that knowledge in political science ought to be grounded
upon a science-based epistemology. The foundation should be empirical reality, that is, that which
can be seen or detected. Concepts should be used to create generalizations; sets of generaliza-
tions constituted theory; and, from theory, generalizations might be taken to use in explanation
and/or prediction. The most precise manipulation of concepts to produce generalizations should
be done using quantitative techniques. That was my belief forty years ago.
As soon as I took this mindset into the field, problems arose. I remember the first interview I
did as part of a survey of cotton farmers in a part of western Tanzania. Helped by a young man
who translated for me, I asked the person to step outside for the questionnaire because I did not
want to “taint” the responses by the presence of other members of the family. It did not take much
time to realize my stupidity. People in that community acted in consultation with others. If I
wanted the attitudes that “predicted” behavior, I would have to get the family together. A few
years later I did a large survey of Ujamaa villagers in several regions of Tanzania and was left
with great doubt about its accuracy, given problems with sampling, language, memory, power
relations, and many other factors. Numbers looked neat and could be analyzed easily, yet I had
doubts about whether they were accurately representing empirical reality. Those doubts grew
over many years.
Today most of my colleagues are very skilled statisticians and they are training our students to
be skilled statisticians. We have never required any methods course except quantitative methods.
The foundation of such work is the data set. Indeed, there is immense satisfaction and excitement
among our students and faculty when a new data set is found. It reminded me of my joy when, as
a student of geology, I found a perfectly preserved fossil. And, the numbers in the data sets are
now referred to as “empirical.” When I went to look at these data sets and compared them with my
knowledge and experiences, I often found a significant discrepancy. A couple of examples: When I
looked at the Polity III democracy scores for South Africa in the 1980s, I discovered that the regime
in that country was very democratic. What nonsense! When I looked at Arthur Banks’s Cross-
National Time-Series Data Archive data for riots, strikes, and demonstrations in India, I found there
were no riots, no strikes, and only four demonstrations in the whole country between 1997 and
- Again, what nonsense! Yet, data sets are the foundation of all quantitative “knowledge.”
When I would point to such nonsense, my colleagues’ response was to challenge me to perfect
the data sets, that is, to make them so they accorded with empirical reality. The more closely I
looked at the protest events, the more I realized the impossibility of the task. Even if I got the