Re-interpreting numbers

The Fearon and Laitin article did a good job at dispelling some of the myths many of us have in regards to civil war and insurgencies. They claim that the usual negative factors, like an authoritarian regime, repression, or inequality have weak correlations to war. However, to what extent can this be true? And should these factors be easily dispelled as irrelevant? One of the things I hate about numerical measurements is its ability to sway many people into thinking certain relationships and correlations are always right, simply because numbers say so.

We also analyzed the Uppsala Conflict Data Program in lecture. One thing that surprised me was their definition between civil war and armed conflict. In their definition, they identified how many battles must be fought, how many people must die, and the duration of the war as standards to legitimately define any sort of conflict. Numbers allow us to measure the intensity and significance of a variable, like war or poverty but it also hinders our ability to understand issues on a general basis. Instead we resort to defining situations based on strict standards and categories. This isn’t a bad thing. It makes it easier for us to interpret information with narrow lenses. However, social scientist should be able to see both sides: one with numbers and the other without.

In my Quantitative Methods class, our professor always told us to be critical of any dataset we analyze. Even though the numbers claim the relationship is weak, you have to investigate the reason why. Moreover, he taught us that numbers are just measurements that try to bring truth to various variables. In no way should we accept them to be absolutely true.

In my Introduction to International Development class, we discussed the significance of the Human Development Index. The HDI seeks to place countries into categories, like highly developed, middle developed, and under developed. This is similar to the way Uppsala categorizes conflicts: armed conflict, civil war, and insurgency. However, we learned that one of the main flaws of datasets and numerical standards are their narrow definition that often leaves it inflexible to other interpretations.

This brings us to the last idea that Professor King says in his lecture on Wednesday, “science isn’t about truth.” We should constantly test theories and numerical findings. The Uppsala datasets and the HDI shouldn’t always be regarded as true but be used as the basis for further understanding. It should always be questioned and re-interpreted. That is the only way definitions change and new standards are made to shed more light on issues that are often more grey than clear cut black and white.