Whenever a problem like gun violence or illegal immigration gains traction in the media, you'll find people misrepresenting the problem. The world is full of uninformed people, but the world is also full of dishonest actors trying to "sell" an idea. This post is the first of several where I will try to help the curious reader defend themselves from the fools and con artists and understand social problems, for real.
Confirmation bias is just the tendency to look for information that confirms what we already know or think. Everyone wants to be right, and the more important the subject is, the more the person wants to continue being right. Information that shows we're right gets more weight than information showing we are wrong. So, with that psychological note in mind, let's look at the outline of a formula for thinking about social problems.
Abusing Statistics - Raw Numbers
A real thinker always gives numbers a context and always uses the right kind of numbers. A serious person will find out what numbers are important and use them. Violent crime in the United States provides an excellent example of this. Suppose you learn that there were 34 homicides in Alaska in 2019 and 1,690 in California in 2019. We'll use murders as an indicator of how dangerous it is to live in each state. It does seem that California is more dangerous than Alaska doesn't it?
Well, 1,690 is a larger number than 34, this fact is not terribly useful. Why? Because there are more people in California. How many more? I made up the murders, but I'll offer 2019 population figures that you can find on Wikipedia.
California - 39.51 million, 1690 homicides
Alaska - 732,000, 34 homicides
Do a little math here, and you'll find that California's population is 53 times larger than Alaska's population. There were 50 times as many homicides in California, not 54 or 55. If California were truly more dangerous, the RATE of homicides, not the raw number, should be higher there. It isn't.
Be careful when someone uses raw numbers, like murders or illegal immigrant arrests, to make a claim about a social problem.
Abusing Statistics - When a Trend Isn't a Trend
A trend is simply a pattern in the data that persists over at least three measurements. That's my own definition, but it serves the purpose of thinking about social problems. Once again, let's consider violent crime to see how this works.
Note that small numbers are easy to distort, while larger numbers are not. Consider a small city where there isn't much violence. That city might have 5 to 10 murders a year. The number varies based on a variety of factors. The rates for armed robbery and rape change in a similar way from year to year as well. A dishonest person or an ignorant one can make us think the city is descending into violence and anarchy with a simple act of deception. Here is how you lie about "the crime problem" in a city:
Step One - Make note of the six murders reported in 2019.
Step Two - Make note of the eight murders reported in 2020.
Step Three - Report that murder is up 33%.
Step Four (optional) - Note a connection between that increase and something you want to complain about, like refugees or a new gun control law.
No one can say if murders are trending up or down based on those two numbers. If we add a third number it might help a little. For example, if there were five murders in 2018, one might think the city is becoming more dangerous.
This thinking is dangerous. Why? You are working with small numbers, so any change looks substantial if someone spins in just right. Note that I mentioned the murder rate going up by 33%. It did, but that doesn't necessarily tell us anything. Look again at the raw numbers. A 33% increase comes from two more murders. What if the city were larger, maybe a little rougher? What if there were 100 homicides in 2019 and 102 in 2020? That's an increase of two murders, but only a two percent increase.
By now you should understand why a simple increase in numbers is not all that interesting. If there is an increase, or a decrease ask this - What is the increase or decrease from?
Abusing Statistics - Comparing Different Things
Let's consider sexual violence in Sweden and in the United States. For sake of simplicity here, we'll focus on one claim that relies on statistics. Someone, probably someone too stupid to know better, claimed that rape is far more common in Sweden than in the United States. Is it though? Are we comparing the same crime or not? To answer the question, you have to know how authorities in both countries classify or define the crime in question.
To make things short, I'll just point out that "rape" in Sweden is not the same crime that the FBI reports on. This fact alone would make it impossible to compare the two countries unless you dive into the data. No one is likely to do that unless they are a sociologist.
Social problems are complex and we don't always agree on what we should measure or how we should compare things. Watch out for dishonest actors who will pick whatever numbers and make whatever comparisons they please. It happens often. It is easy to miss this when you want the other person to be right.
Abusing Logic- Cause and Effect
Did you know that ice cream causes crime? Really! Sociologists have found that as ice cream sales increase, burglaries tend to increase. Of course, this is nonsense, but it sounds cool doesn't it? In truth, burglary tends to be more common in the summer, when ice cream sales tend to be higher. Summer drives both things.
Establishing that X causes Y is a complex undertaking. It cannot be simplified by using common sense. Consider the case of a new gun control law that makes it illegal for most people to carry concealed handguns on them or in their vehicles. The law passes, and crime goes up. Our imaginary state saw 3000 violent crimes in 2018 and 3300 in 2020, the first full year of the tougher law. Well, that increase in 300 crimes is a 10% increase. Doesn't it seem like that gun control law was a dumb idea?
No, it doesn't. Ignore the propaganda and reflect on what we know. We know the crime rate went up. We know gun control got tougher in the year before the crime rate went up. That isn't enough unless you need it to be (re-read that paragraph on confirmation bias) because reality is complicated.
Note that I only used two years' worth of crime data. This is a deliberate act of deception that many activists use. Note also that we know nothing about other events in the city. Maybe the population has been growing for a few years. Maybe the economy began to shrink in 2019 and many people were becoming impoverished.
Abusing Logic - The Single Cause
Why are some people poor and others are not? Why is there more crime in some cities? Whatever overly simplistic answer there is, there is a vlogger, podcaster, or television "analyst" peddling that answer. Immorality causes crime. People are poor because they are lazy. Black people are poorer than While people, because Blacks don't want to (or cannot) do any better.
Pay attention to the explanation you hear for a social problem. This is even more important when you want your source to be correct. If you are conservative and hear someone claim the real cause of crime is immorality, you might be tempted to believe. Once you make the snap decision to accepts this Immmorality Theory (not a real theory), you will look for evidence to support it. That's confirmation bias.
Asking the Right Questions
When pundits, YouTubers, television commentators, and Facebook trolls post anything on a social problem, evaluate what you see with this formula:
- Are they using raw numbers and rates correctly?
- Are they really presenting a trend here or just two numbers that happen to be recent?
- Are they comparing apples to apples or apples to basketballs? How do you know?
- Are they presenting two numbers that have something to do with each other, or do they just look like they go together?
- Do I just want this person to be right, or are they presenting valid figures and conclusions?
- Does the explanation seem to be oversimplified?
- Are they using anecdotes/stories instead of data to make their case?
If the answer to all applicable questions is "yes" or "probably" then that article or editorial or whatever is at least attempting to present the issue fairly.
What do you think? Should I add something to this formula?
Next time, I'll go over some of the same topics but I will use them to explore solutions to social problems.
Comments
Post a Comment