Lying Statisticians are the Real Terrorists

White males commit more terrorism than any other group. Everybody knows this, right? If you try to debate terrorism on social media, the Smug Leftist will play his trump card: "ACKSHUALLY, statistics prove that WHITE MALES commit the most terrorism!"

This may not seem right to you, despite the fact that there are indeed headlines proclaiming that "fact." But you may be at a loss for how to argue the point, since numbers don't lie, right? Well, I'm here to help.

I spent several years of my life in government agencies' data analysis departments. I will tell you there is often political pressure on analysts to "prove" a certain point with data, and there are plenty of ways to manipulate or creatively interpret data to make it seem to point one way or another. Dishonest (or possibly dumb) journalists will happily tout a headline that a "study says..." something that they want to be true. Most people will simply absorb the headline, assume there is some truth to it, and move on. 

I'll try to give a quick summary of some of the common ways that numbers can lie:

1. GARBAGE IN, GARBAGE OUT - This is a common phrase in  the world of statistical analysis. It means that if the data collection is no good, then the resulting data analysis will be no good. You must ask how the data was collected and by whom. Was it collected by trained, conscientious, and non-biased workers? Often the answer is no. It is wise to be skeptical of information collected by workers who have inconsistent training, a disinterest in data collection, or a motivation to skew data to secure agency funding or cover their own failures. Also, if you see the phrase "THERE IS NO EVIDENCE OF X" , you may want to ask whether there is no evidence because the data was never collected!

Self-reported data is also notoriously unreliable. People can have highly inaccurate self-perceptions. They may make errors when responding to surveys, or they may choose to fudge or even outright lie in order to achieve the result they prefer or avoid embarrassment. Compared to carefully conducted scientific analysis, such information can be considered "garbage."

2. TOTAL NUMBER VS PER CAPITA. This is an especially useful one with regards to the "White people are the worst" arguments. Because Whites are a majority of the United States population, we are likely to have "more" of any given problem if you simply count the numbers. The "Whites are the worst" argument often falls apart when you look at the numbers as a percentage of the group population. This may seem obvious, but it is easy to slip past most casual readers of mainstream media, especially if it confirms what they wish to believe.

3. APPLES TO ORANGES. A few years ago some media ran reports that children raised by gay couples are happier than those raised by conventional families. This is an example of the "apples to oranges" deception. The conclusion was based on a comparison from two separate studies: a long-term, scientific study of children in conventional homes versus a simple survey completed by gay parents who volunteered to be questioned. The data was collected from two separate sources, one with a much more reliable methodology than the other. They are simply not comparable, at least not if you're trying to be honest and accurate.

4. FUDGING DEFINITIONS AND CATEGORIES: A tremendous amount of mischief can be caused by this one.

Most people know that unemployment statistics have been manipulated by fudging definitions. When a person has been unemployed so long that they have fallen into despair and given up the search, they are no longer technically a job-seeker, and therefore not "unemployed" for statistical purposes. So, excluding the long-term unemployed from calculations may make employment statistics appear to improve, even though no positive change has occurred.

This kind of fudging is critical for the "real terrorists" ruse. Most of us think of mass-casualty shootings or explosions when we hear the word "terrorist." But does the study cited count other things as terrorism? Arson, vandalism, unarmed one-on-one scuffles? Most mainstream articles don't provide that level of detail, and most people don't think to ask. If a guy who happens to be in an all-white biker club beats someone in a bar fight, would that be counted as "white supremacist violence?" If a member of an Aryan prison gang knifes another inmate, would that count? These are examples of criminal behaviour, of course, but not what one normally thinks of as terrorism.

Also, what is counted as an extremist group? If you are reading this site, you probably know that the SPLC definitions are extremely biased, and elastic to the point of being useless. They may include completely peaceful advocacy groups and orthodox religious sects. When an article cites a growth in "extremism" it may sound alarming, but it could just mean that more people are supporting the Family Research Center or other organizations that the study's sponsor doesn't like. It could even simply mean that they have expanded the list of "hate groups" to include new things. 

You may even want to question the definition of White, as some studies include hispanics or people of middle-eastern descent in the "White" category.

5. IN WHAT UNIVERSE? One of the most important parts of any study is choosing the sample population. Who are you asking? Is the study based on a representative sample of the relevant population? For example, if the findings are based solely on the study of elderly Scandanavian vegetarian nuns, it may not be applicable to other groups. More realistically, If the sample is self-selected (for example, if it is conducted with people answering an ad asking them to participate in the study), the sample is not random, therefore introducing bias into the results. 

Leading up to the 2016 presidential election, many polls of "likely voters" showed that Hillary Clinton was consistently ahead. The trick was in the sample selection. Manipulating the proportion of Republican, Democrat, and Independent voters in the sample affected final results, and in a way that casual observers did not catch.

6. MISUSE OF OUTLIERS: A outlier is an extreme example that lies outside of the normal range of something. Including the "outlier" information in an average can be very misleading. For example, averaging the seven-figure salary of a CEO with the salaries of the regular employees would create a misleadingly high impression of the typical company salary. 

This tactic should be kept in mind when people spout statistics about gun crime. The national gun crime rate is high compared to many other countries, but if you take out the outliers, that is, the top 10 or so most gang-infested cities, the overall rate goes down substantially. Including the numbers from the outliers, in this case the dysfunctional urban areas, makes the overall national rate much higher. This misleading impression is helpful to those who would like to imply that a high rate of gun crime is nationwide problem.

7: PUSH POLLS: "Would you support a mandate that millions of illegal immigrants be ground into hamburger and fed to alligators, or would you prefer allowing those who meet certain reasonable requirements to stay in the United States?" SHOCK POLL! 99% of Americans support undocumented migrants becoming citizens!

This is an exaggeration, but you get the idea. Don't accept "survey says" pronouncements at face value. If you think the results are suspicious, dig deeper to find out exactly what questions were asked and to whom. 

There are many, many more tricks that can be employed by those who want to influence the public with biased "facts." It's never easy to discern the truth in our information-overloaded world, but hopefully this article can give you an idea of some questions to ask when the facts seem fishy.