Russian artillery fire in Chechnya from 2000 to 2005 - English

The data we are going to study deals with the bombings launched by the Russian army against the Chechens between the years 2000 and 2005.
Between the years 1999 and 2000, Russia was engaged in a \”total\” war against Chechnya: \”the second Chechen war\”. The war between these two countries began on 26 August 1999 and ended on 6 February 2000 with the capture of Grozny by the Russian army. However, as a result of this war, Russian forces continued to carry out numerous indiscriminate bombings against Grozny, the capital of Chechnya, as well as several villages in its surroundings. The aim of these indiscriminate bombardments was, according to the Russian leadership, to avoid independence revolts after the siege of the capital. However, these indiscriminate bombings caused the death of many civilians and destroyed most of the city. It was the most violent conflict in Europe after the Second World War. Even though the Chechen rebels were the main target of the Russian army, they represented only a minority of the population, while the bombing affected the entire population. It is therefore legitimate to question the effectiveness of these bombings against the Chechen rebels.
In this research, we analyze the relationship between indiscriminate violence and insurgent attacks using data about Russian artillery in Chechnya from 2000 to 2005. Some believe that indiscriminate violence increases insurgent attacks by creating more cooperative relationships between citizens and insurgents. Others believe that indiscriminate violence can be effective in suppressing insurgents\’ activities.
It is interesting to question that, when Russia indiscriminately bombards a Chechen village does that village tend to retaliate or not? Does it increase or decrease the number of revolts?
In other words, what is the correlation between Russian bombing and insurgent attacks?

The data set that we will study in order to answer these questions contains 6 different variables.
The first variable of our data set is the name of the bombed villages, so it is a nominal quantitative variable since it cannot be hierarchical, no value is greater than another.
The next two variables are the \”Grozny\” variable and the \”fire\” variable. The former determines whether the village is inside Grozny (1) or outside Grozny (0). The second variable (\”fire\”) determines whether the village was bombed (1) or not (2). These are therefore two dichotomous qualitative variables because they cannot be prioritized and have only two possible choices.
Then, the last three variables of our data set are the

variable \”deaths\” which designates the number of deaths following the bombings, the variable \”preattack\” which designates the number of insurgent attacks before the Russian bombings, and the variable \”postattack\” which designates the number of insurgent attacks after the Russian bombings. The values of these variables can only take on entire numerical values, and the transition from one value to another is discontinuous. They are therefore discrete quantitative variables.
This is why we are going to use the calculation of median, average, as well as the calculation of percentages in order to understand the correlation which exists between the Russian bombardments and the attacks of the Chechen insurgents.

First of all, we are going to check how many villages were fired by Russians and how many villages were not fired to have an idea of the proportion.

There are 318 observations. In these observations there are 159 that were fired by Russians, and 159 that were not. There are as many villages that have been fired than villages that not have been fired by Russians.
Then, in order to try to understand which correlation exist between the Russian bombardments and the attacks of the Chechen insurgents, we are going to create a subset of the villages which are in Grozny and out of Grozny in order to analyze the number of deaths.

In Grozny, Russian attacks have led to the death of 3.71 people in average with a median of 3 deaths. On the other hand, outside Grozny the mean is 1.57 deaths and the median of 0 death.
Russian attacks result in a greater number of deaths (more than double) in Grozny than outside, in terms of both mean and median.
Moreover, in order to answer the questions we are asking ourselves, we need to compare the mean number of insurgent attacks after Russians fired in the villages hit by Russians to the average number of insurgent attacks after Russians fired in the villages not hit by Russians. We need to address this kind of methods to be suitable from a data science point of view because it is thanks to these that we will be able to interpret our data set.

Thanks to these methods we can see that in the villages hit by Russia the mean number of insurgent attacks after Russians fired on a village is 1.5. Whereas, in the villages that were not hit by Russia the average number of insurgent attacks after Russians fired is 2.05. However, I would not conclude that indiscriminate violence reduces the number of insurgent attacks because we do not know yet the situation before some villages got fired by Russia. We cannot state anything on the impact of the indiscriminate violence on the evolution of the number of insurgent attacks because we have not studied yet the situation before.
So, we are now going to see if before Russia attacked villages, both type of villages had a similar number of insurgent attacks or not. With these results we will be able to give meaning to the results we have just obtained.

Accéder à l'échantillon complet