I listened to an interesting argument on the radio recently. It emphasized the understanding of statistics most people lack. This issue is important because statistics are used, misused, abused, and prostituted to justify and garner support for political causes. Among those are laws and bills pertaining to self-defense, such as “gun control” (more appropriately titled “civilian disarmament” or “firearms prohibition”), knife restrictions, regulations on martial arts training, and liability guidelines.
Citing a recent study, one participant in the radio debate claimed that drivers of mid-sized sport utility vehicles (SUVs) are “nine times more likely to die” (from rollover accidents) than drivers of passenger cars. Over and over again, this woman repeated the statement, “You’re nine times more likely to die if you drive one of these SUVs. Nine times!” Her opponents countered weakly that the study was “nonsense” and that, intuitively, drivers are safer when in larger vehicles. They pronounced it absurd that our government “wants to make the bigger vehicles less safe to make everyone die equally.”
The flaw in the entire argument, and which none of the participants could see, is readily apparent to anyone who understands statistics: the traffic data analyzed was a list of all automobile fatalities for a given period. Think about that. All the accidents were fatal. This means that to gain a realistic appreciation for your odds of suffering a terminal vehicular accident, you must examine all traffic accidents that did not end in death and figure that into your calculations. You must ask yourself, “If I am in an accident of any kind, what are my chances of being killed? How many accidents did not end in death for drivers of SUVs versus how many did?” Only then can you say with any confidence what might be your “chances of dying” relative to drivers of passenger cars.
The same “reasoning” was used in one of those famous “studies” that is now accepted as folklore truth among firearms prohibitionists – groups that want to make it more difficult for you to protect your family and your person. These organizations want to eliminate your right to arm yourself. The figures cited vary, but usually conclude with the dire pronouncement, “A gun in your home is X times more likely to kill a family member than someone else,” or “If you own a gun your chances of being killed increase by X times.”
The questions you should always ask yourself when confronted with a given set of statistics are these:
- How was the sample taken?
- How big was the sample? Was it taken randomly? Is it representative of the population, or is it skewed?
- Was important context left out while taking the sample?
- Are the conclusions drawn logical, based on that sample?
Looking at the issue of context and considering the gun violence propaganda, we will – in investigating the sample – discover that the data was taken from an analysis of gun-owner homes in which a violent act occurred. The missing context? All the homes occupied by law-abiding gun owners in which no violence occurred have been excluded from consideration, resulting in a distorted perception of the probabilities involved. To truly know your chances of experiencing harm as a gun owner, you must examine a representative sample of gun owners to extrapolate your own chances. You cannot simply look at only those homes in which the result (for which the propagandists are searching) has already happened.
Using the flawed statistical methodology of propagandists, we could calculate your chances of being killed by mortar fire by examining those third-world nations in which wars are occurring at this time. Of course, if you live in a nation in which war is not occurring but we excluded nations like that from our study, our statistics are meaningless and skewed.
Statistics are not hard to interpret if you have the relevant details of how the statistics were produced. Anyone who can’t or won’t provide references for those details is lax – or hiding something. Do not, however, simply give up and proclaim that all statistics are meaningless. In context, statistics are relatively difficult to distort – without being painfully obvious about it.
Do not accept at face value the statistical slogans people throw at you in supporting their opinions. Demand the sampling context and apply logic to each scenario.
Be mindful of this when using statistics yourself and you will have a powerful persuasive tool at your disposal.
How big was the sample? Was it taken randomly? Is it representative of the population, or is it skewed?
Was important context omitted in taking the sample?
Are the conclusions drawn logical, based on that sample?