Home » AMERICAS » On the Ethical use of Statistics and Attack Drones

On the Ethical use of Statistics and Attack Drones

This article is in response to Daan’s article “Condone the Drone”. The objective of this response is twofold: first, it aims to promote a critical assessment of statistics and how they are used in political argumentation; second, it aims to re-frame the debate which was presented by the author of the original article. For the purpose of coherence, I suggest that this article is read in tandem with the article which it critiques and the statistics which the targeted article uses. Now that this is cleared up, let us begin.

Seeing as political science is mainly a positive science, based on empirical observation and the comparative analysis of the accumulated data, statistics obviously play a huge part in confirming a central thesis. In political argumentation, the misuse of statistics is widespread: on higher levels, it is due to purposeful instrumentation (most politicians are adept at this type of misuse); on lower (academic) levels, it is usually due to lack of scrutiny in the comparison of statistics.

The Ethics of Statistics Use

In “Condone the Drone!”, the author uses statistics to confirm the validity of his central thesis. Unfortunately, the statistics are misused and end up creating misinformation; this leads the reader to believing in the validity of the author’s central thesis, when in reality, the thesis itself rests un-supported.

The author commences the article by citing a statistic, presented by Peter Bergen, which is meant to support his thesis and which states that drone-based civilian casualties “ hovers at around 16% and is decreasing further.” The author then quotes another statistic by the Bureau of Investigative Journalism which proves a counterpoint to his thesis (it states that drones are not as safe as Bergen makes them out to be), and he aims to refute it. The Bureau’s article is a critique of the statistical method used by Bergen; it simply proves that Bergen’s statistics employ questionable methodology.

Instead of criticizing the Bureau’s own methodology, the author chooses to refute the Bureau’s statistics with the statistics from a UN study on air-based civilian casualties in Afghanistan. The author shows that the UN statistics report that “…civilian deaths in Afghanistan air strikes fell by 42% in 2012, the same year in which drones became a significant presence in Afghan airspace.”

This is precisely where the author’s method becomes questionable. Here’s why:

First, the UN statistic only states that “…civilian deaths in Afghanistan air strikes fell by 42% in 2012”. It makes no distinction between drone kills and plane kills, as it lumps both into the same number. Without the recorded number of drone kills in Afghanistan, the UN statistics mean nothing.

Seeing as the UN article is the last leg which the author bases his statistical argument on, let me put a few things into perspective. The author says:

“Let us first address the statistics employed by the Bureau for Investigative Journalism; the study by the BIJ argues that collateral damage from drone attacks is high, but contradicts a United Nations study that reports civilian deaths in Afghanistan air strikes fell by 42% in 2012, the same year in which drones became a significant presence in Afghan airspace.

By saying that 2012 was the year where drones “became a significant presence in Afghan airspace”, the author aims to strengthen the almost non-existent correlation found in the UN stats by simply inferring that “more drones + less casualties=proof that drones are safer.” The problem is that the UN article has no mention of “drones becoming a significant presence is Afghan airspace” in its statistics. Seeing as “…drones becoming a significant presence in Afghan airspace.” is a statement which must also be backed by precise data (which cannot be accurately provided, as the number of deployed drones is top secret), his line of reasoning hits a wall.

Let me suggest an alternative.

Let us measure “significance of drone presence” by a statistic that we can measure: frequency of drone attacks. The reason for this is simple: the US could be employing 50 or 5000 drones in Afghanistan; what matters is how many times they attack. Their presence, in relation to civilian casualties, must be measured by the number of attacks.

Here’s a screengrab taken from the Wikipedia page for drone attacks in Pakistani-Aghani border (where most of them take place). The stats found in the screenshot were compiled by the New America Foundation, which is the think tank Peter Bergen (whom the author cites) works for. They are then contrasted with the Bureau’s stats (which are rather similar).

Do you notice anything?

The reason that the amount of civilian aerial-based casualties has decreased in 2012 is not because drones are safer than manned planes: it’s because there was a significant decrease in drone strikes in 2012, compared to 2009-2011.

It seems that 2012 was the opposite of what the author stated: it is a year of decreased drone presence. Because of this, his use of statistics is mis-informative. By using statistics which do not validate his thesis, followed by the inventing of a statistic (the increased presence of drones in 2012) to hold it all together, the author is unethical in his methodology.

The Ethics of Warfare

In “Condone the Drone!”, the author frames the ethical debate on drone strikes around the dichotomy of “simulation/reality”; he presents the “leftist” criticism of drone strikes as centered around the idea that allowing unmanned vehicles to kill people is unethical. He then (mis)uses statistics to try and demonstrate that drone strikes are safer to what he assumes is the “leftist” alternative: using manned jets. In reality, this is a false dichotomy. What is at stake is not how enemies should be killed in war; getting blown up by a plane and getting blown up by a drone amounts to the same result. What is truly at stake is how drones are used.

The Obama administration authorizes the use of signature strikes in drone warfare. A signature strike is  an attack where the drone operator is allowed to strike without having to properly confirm whether his/her target is an actual threat. This means that a drone operator can attack a target which might or might not be an enemy, and wouldn’t have to suffer any consequences if his target was actually a civilian.

The problem of killing without confirmation of threat is one which always plagued warfare. The drone was incorporated into the US air force arsenal in order to allow for a better assessment of threats to take place; the drone’s ability to hover and gather reconnaissance is what is supposed to allow this. The goal of the attack drone is to have a weapon which can better differentiate between enemy and civilian. The signature strike program completely negates all of these newfound improvements, and transforms the drone into a soldier in the sky. Now that drone operators can function on the same basis as ground soldiers (deciding who lives and dies without proper evidence), what the US military has created is a weapon which loses all newfound reliability while gaining even more of a destructive capability.

The signature strike puts the drone operator at the forefront of operational decision-making; the environment which the drone operator works in makes this even more problematic. Long shifts, coupled with the desensitization that comes with viewing human beings as simple IR signatures on a screen can only worsen the operator’s ability to critically assess which target is a genuine threat or not. A comprehensive and standardized threat-assessment procedure is what is used to counter the effects of the operator’s environment; the second it is removed, the lives of civilians become dependent on the external factors which the drone operator might not be aware of.

As you can see, it is with signature strikes that the dichotomy of “simulation/reality” becomes the center of the debate. If signature strikes are forbidden, and measures are taken to assure that the operator’s environment/subjectivity don’t have say in who lives and who dies, the debate can evolve into larger and more complex one. As it stands, the ethical inquiry should follow this primary question: “If there is not enough evidence to confirm whether the target is an enemy or not, should that target still be eliminated?”

This question does not only pertain to warfare; it can be found in every domain where individuals have the choice to determine who can live or die. If a judge shouldn’t put an inmate on death row without precise evidence of his crimes, or if a cop shouldn’t shoot unless he is sure that the perpetrator presents a threat to the lives of himself/civilians, then I do not understand why drone operators should not be put under the same amount of scrutiny. Justice transcends contingent barriers such as these; if signature strikes continue to take place while American law enforcement officials have to follow more “humane” (prudent) procedures before killing, it simply shows that the United States Government does not accord the same amount of importance to the lives of foreign nationals than it does to its own citizens. Many will say that this is expected of any country; either way, it doesn’t make it right.

If we want to debate about the ethics of aerial warfare, it is pointless to argue whether killing a civilian on a computer screen is more or less ethical than killing a civilian “in real life”. What must be discussed are the ethical implications of allowing operators to designate targets without having evidence of a threat or not. The fact that the author  does not mention signature strikes, especially when they have been increasingly mentioned in mainstream news, is quite disappointing. A real discussion could have been had, but I feel that this was not the goal of the original article.

– L. David


Featured photo: Attribution  Abode of Chaos, Creative Commons, Flickr

About Guest Writer

Check Also

Running for CSU? Runner Beware

By Eddy Kara The Concordia Student Union (CSU) holds annual elections for their executives and ...

One comment

  1. Great piece, maybe too focused on statistics but I guess this issue was important to point out.

Leave a Reply

Your email address will not be published. Required fields are marked *