Shrikant speaks the truth...

My lifelong memories are false. Riots in India are not a national phenomenon; they are highly localized. In a country that remains mainly rural, religious violence is a chiefly urban problem: the countryside accounts for under 4 percent of all riot-related deaths. Using data from 1950 to 1995, Varshney shows that riots are concentrated in just four of India's 28 states. On a per capita basis, the worst states are Gujarat, Bihar and Maharashtra.

I agree shrikant. In a vast country, population, socialogical mixture, geo-politcal circumstances and the poverty some facts can be validly quoted by Ashutosh Varshney in his research. These definitley seems to be virtual and totally based on case studies. I will say during my life in India for about 22 years. I never saw any riot, just heard of them (I donn want to see also).
The inter ethnic turbulance put more energy in the bad cup. We need to understand the need based social structure and also sharing partners for the resources. The place where the population is growing irrespective of speed of the social living standards. The people just leaving their capsules of joint families and learning the need of a short family. We are still attached with lot of superstitious faiths. Seventy percent of whom has not known the Internet.

The study definitely seems like "sophistication of cutting edge Western scholarship with the empirical richness of micro-level cases studies from India".