Sunday, November 6, 2016

Responsible Statistics?

"The possession of knowledge carries an ethical responsibility."

It has become increasingly common for emotional appeals to contain statistical reasoning. When arguing that there is an issue between the police community and the black community, people vaguely reach for statistics like "black unarmed men are killed at three times the rate of whites and other minorities." When arguing that there is still unjust sexism towards women today, statistics are used like how 1 in 3 woman are sexually assaulted by a partner during adulthood, or how women are ten times more likely to be victimized than men. This statistical name-dropping is binding individuals up in numbers in an effort to make an impact on the ethical conscious of the audience.

This act of processing victims in the form of numbers, shuffling them like all they are is data, seems to lead to an ethical dilemma: to represent individual tragedies by a number feels like it's dehumanizing the injustice. But that isn't what is meant by the speaker who crutches on statistical appeals; statistical reasoning is used to prove the validity of an issue.

It's an odd quark in the system because mathematics isn't usually associated with empathy or ethics.

The interesting thing about any statistical evidence is that there is undoubtedly large amounts of statistical bias floating around in this field. Statistical claims are more or less simply probabilities. Data is collected and processed, but unlike the theories Laplace's scientific determinism, there is no way to access all the data relating, especially when it's data relating to human relations. These weaknesses in statistics aren't common knowledge. The field of mathematics is viewed entirely objectively, and seen as something without loopholes or shortcomings: more or less, it is seen as an undeniable fountain of knowledge.

Statistical reasoning, when used to appeal to emotional reasoning, can easily become dangerous. Faulty statistics, or even justified statistics, can be twisted and bent to serve specific purposes in an argument. This deceitfulness can be said for most anything though. Evidence can always be cut and pasted to be interpreted any multitude of ways.

Ethics can get muddled, but if ethical responsibility means maintaining honesty and transparency, statistics easily fails. It is agreeable that it is the responsibility of citizens to view their fellow compatriots as real, valid, and not bound up in a number, sure. But furthermore, into the mathematical realm of these numbers, it is the responsibility, whether ethically beneficial or not, of people to question mathematical reasoning. Because having an ethical conscious extends beyond emotions and moral compasses, especially when the field of mathematics overlaps the field of ethics.

1 comment:

  1. Another really thoughtful post, though it gets a little muddled at the end. I love your point about the ethics of dehumanization, and of course statistics are bandied about so irresponsibly now that they are hardly trusted at all. It's a tough place to be: so many people have misused statistics, that statistics are automatically considered suspect. Indeed, I often feel that, for certain types of people, I come across as LESS TRUSTWORTHY if I start to feed numbers into the system.

    That said, statistical reasoning IS just that: reasoning. And I find that playing the numbers game is an excellent check on my own miniscule worldview. Just because, "I've never disrespected a woman," doesn't mean that women are not, on the whole, the victims of disrespect. And yet, to say "women" is to categorize based on demographic, which is to introduce a whole bunch of assumptions into your statistical model.

    I wonder if there is an argument for "Responsible Statistics." This would look like Nate Silver's blog: Carefully qualifying your claims and keeping the uncertainty of the system at the center of all of your conversations.

    LOVE the comic.