Sex and Violence

The article, Jane Austen and the case for extreme porn, from the UK's TimesOnline got me thinking. Not about porn. Not about sex or violent sexual acts or violence against women or even Jane Austen. It got me thinking about statistics. Yes, statistics.

The article states that the proposed law to ban possession of violent pornography, "followed a long campaign by a mother whose daughter was killed in 2003 by a man who was said to have been a obsessive viewer of violent porn sites." And that's where my mind started going back to my stats and experimental design courses from back in the day when I was a psych major. You see, I've never been convinced by people that say that those who consume violent media or violent pornography are inspired by that media to go and commit acts of violence. I have yet to be convinced there is a causal relationship. People then point out the frequency of violent offenders possessing extreme pornography or being fans of violent video games, music, or television. They say that this means there is plainly a causal relationship between consumption of violent pornography and committing acts of violence against women. It doesn't.

Showing a causal relationship is extremely difficult. Statistics and experimentation can reveal a causal relationship with some effort, but that takes quite some time and the burden of proof is rather high. The frequency of two elements occurring together does not mean that one causes the other. The example from the first day of my Stats for Psychology class: "In the summer, ice cream purchases and the murder rate both go up, therefore eating ice cream causes people to commit murder." Well, that makes no sense, you say. Ice cream and murder aren't related at all. Eating ice cream doesn't make people go kill other people. That's right, it doesn't, so there must be another factor in there, one of these things isn't causing the other, their must be a related causation. Like, perhaps, that it's summer time and people eat more ice cream because it's hot, and more murders are committed for some other reason, but it also has to do with it being summer, not the ice cream consumption. Even those suppositions, however, aren't really accurate ones to make. We'd have to do a complex study on murder and ice cream sales to really get a sense of what is driving these numbers upward.

My problem with bans like the ones proposed goes beyond my feeling that it is restricting personal liberties, and beyond that it is presuming the guilt of anyone consuming the violent pornographic materials. My problem is they reveal a misunderstanding of statistics that runs rampant through, at the very least, western societies. We base many decisions on statistics and yet, so few of us understand enough of the basic principles behind reading statistics. Let's review a couple of brief rules to live by when reviewing statistics First, statistics rarely, if ever, show causation. Second, statistics are based on a set of collected data. Always consider the source of the data and how the data was collected. Third, statistics don't lie, but they can be used to lie.

Statistics show the relationship between two measured data points. That's all they really do. All the law in question is based on is that there happens to be a relationship between consumption of violent pornography and committing violence against women. People immediately jump to that the relationship must mean that one is causing the other. Not so, and while I will admit that there could in fact be a causal relationship I will argue that that relationship has not been shown. It is much more likely, based on the information available, that the relationship is that people who are prone to committing violence against women may also enjoy viewing violent pornography.

Just last week my girlfriend and I were discussing statistics and experimental design and that I feel that these need to be subjects taught to everyone, they should be basic High School courses. Some people look at me a bit funny when I say that, but understanding how to read statistics and understanding what an appropriate experimental design is are critical thinking skills that will help people process information presented to them on a daily basis. Newspapers are filled with statistics and reports on the results of experiments. Understanding the building blocks and principles of experimentation and statistics will help people understand the information presented to them every day. One of the arguments I remember for learning geometry in High School was that I would need it some day. I'll need to know how to figure out the area of a square or triangle because maybe I'll be doing gardening or something like that. I don't do a whole lot of garden building. The girlfriend and I got into a bit of a debate over this, she liked geometry but algebra never really clicked for her, and I'm the opposite, so my insistence is that people should be learning stats in High School rather than geometry, and she's arguing that for a lot of people algebra makes no sense, but they can follow geometry because it is grounded in concrete objects. Maybe stats doesn't have to replace either subject, maybe it could be just an additional subject, but it needs to be something everyone is exposed to, and I'm not talking about having to figure out what the Standard Deviation is, or any of the other math stuff that happens with stats, I'm just talking the basic principles. Understanding the basics that will help you read and critically evaluate data presented to you.

Experimental design is similarly important. We're always presented with stats based on surveys, and a lot of us take those stats at face value. Most of us are not equipped with the questions that will allow us to evaluate the basis for the stats presented to us. Learning about Experimental Design will equip people with the tools to question and understand the quality of the data being presented. Since taking a class in Experimental Design I have taken most studies and surveys with a grain of salt. When looking at a survey you should always ask: How many people were surveyed? How were those people selected? How did the people respond to the survey? Was it over the phone, filling out a sheet of paper, or a face-to-face interview? Was the survey conducted the same way for all participants? Is the group selected an accurate representation of the population the statistics are being used to describe? Who asked the questions? What questions were asked? How were the questions asked? Being able to at least ask these questions to further your understanding of the data will help you understand the difference between quality information that you can use to your benefit, or whether the data before you is based on a biased or at best collected by a poorly designed tool.

Understanding Experimental Design and Statistics are fundamental skills in our information age. Fundamental skills that very few people are taught. Because of that we get laws proposed based on assumptions of causation where none may exist. People are having liberties curtailed because of a misunderstanding of data and that's wrong.