Readers’ comments: Important report on increase in communal riots has a data flaw

The 84% rise in communal riots in 2024 in India, compared to 2023, is a very significant finding (“India saw 84% rise in communal riots in 2024: Report”). I may have spotted an important methodological issue, though. The report by the Centre for Study of Society and Secularism, which (quite rightly) uses newspapers as the source of data, does not include “communal riots that are either not reported or are reported in other editions of these newspapers”.

It is, indeed, true that newspapers do not always report riots that did not lead to fatalities. For a scholar, this is “missing data” – a common nuisance that can adversely affect their findings. Given that quite a few riots likely went unreported in both 2023 and 2024, or either year, a comparison of riot occurrences is unlikely to be fruitful or accurate. For example, if 2023 had as many riots (many unreported) as in 2024, it tells us a very different story – that nothing changed between the two years.

A better method would involve counting fatalities rather than the number of riots. Newspapers will always report riots with fatalities, even if the death toll is only one. This better-quality data would reduce the bias arising from missing data…

Read more