Statistical Fallacies and Debating

 

 

 

I once took a statistics class during my undergrad years in college. The professor opened the class on the first day with introductions and an ice breaker question, "Why are you taking statistics?" The truth was that I was taking stats as a math option for my major. If a student was a social sciences or humanities major, statistics was allowed in place of a math class. And, when the intro time fell to me, I quoted Mark Twain as my reason for taking the class:

 

 “There are three kinds of lies: lies, damned lies, and statistics.”

 

Needless to say, my relationship with my professor was rocky all semester. Go me!

 

I have noticed however that in heated debates that draw on numbers in any way, statistics are often not trusted, as if Twain was right. Gut instinct, personal observations, and presuppositions are often weightier for people than deep research. So in these sorts of arguments, statistical fallacies are often employed unwittingly, naively, or ignorantly by detractors. 

In my recent debates with friends and family over important issues, such as climate, covid, gas prices, immigration, and so forth, I have seen three sorts of logical fallacies employed, and two are specifically statistical fallacies.

Anecdotal Fallacy

First, there is the anecdotal fallacy. An anecdote is a personal observation or story. It relies on perception as the chief reason for an argument. Case in point, one of my most draining debates recently had to do with President Biden and gas. My opponent posited, as many in the GOP are presently doing, that because Biden withdrew oil exploratory and fracking permits from federal lands, and because he nixed the Keystone Pipeline, our present gas prices have risen.  

These arguments against Biden began shortly after he took office in January of 2021, when he cancelled the permitting of the Keystone Pipeline expansion. This had no direct impact on gas prices, since the line was not yet producing, being a construction project only. It is possible that energy speculators affected prices in anticipation of Biden's climate platform, but to have affected the prices so dramatically and globally over one non-producing pipeline is highly unlikely. And the blame for gas prices began in January, which was perhaps too sudden. Biden seems to be a mere scapegoat. 

Biden had also banned fracking and oil exploration on federal lands. But, was that enough to account for a global rise in gas prices, which interestingly paralleled inflation regarding other commodities? Unlikely. Gasbuddy.com, which tracks gas trends and statistics, was even cited in a chart in a USA Today article showing that in 2020, gas prices suddenly plummeted with the advent of covid-19, but by April/May began to rise again, and by December/January, the rise coincided  with covid vaccine. And this was global. 

 

Economists show that what had happened is that with covid, people weren't traveling amidst mitigation and shut down efforts. To sell off the surplus, prices fell. People also lost jobs in the energy industry. As mitigation efforts began to work, by May production increased some and prices rose. In the Fall of 2020, things stabled out some. But, when the vaccine hit the market, more people resumed a sense of normality. Demand for gas and other goods spurred a drastic increase in prices, because the supply was too little and closed production facilities now scrambled to meet the demand. This meant rehiring people, purchasing supplies from an impacted supply chain, and ramping up production. It was a catch-up and it was costly... These costs were applied to the prices we pay today. It is a matter of supply and demand. 

But to many, gas price increases seemed to be connected to Biden by the coincidental timing of his inauguration. This is the observation. This is the anecdotal evidence. At a glance, it seems enough to hypothesize a relationship between the two factors, Biden and prices. But the history, statistics, and exogenous factors don't seem to support this hypothesis, so to use it as an absolute fact is merely circumstantial. 

 

Another example is illegal immigration. Many people assume that with the current increase in numbers of migrants at the border and those illegally crossing, that Biden's policies have suddenly become "open border" and highly liberal. Yet, a check of Biden's official policies remain within the restrictive nature of our immigration act, have attempted to dissuade illegal crossings, and have kept many of Trump's policies intact. The few areas Biden has changed, such as halting wall construction and attempting to deal with unaccompanied minors and families more compassionately, don't actually answer for the numbers coming and don't account for the fact that the vast majority of crossers are deported in line with prior policy. 

 

It is entirely possible that migrants expected Biden to be more accommodating than Trump and so they came in droves. But to say that Biden invited these people and changed policies to allow illegal immigrants into the country is a leap based on anecdotal connections only. People see the masses of migrants and presume that Biden has changed the system to allow them in. It is a conclusion based on superficial observations, not the reality of policy or practice. It is a logical fallacy based on perception only.     

 

Simpson's Paradox

 

Another statistical fallacy is called Simpson's paradox, which mistakenly looks at aggregate information or the whole, and presents the information to make a case, without considering independent factors or variables within categories. Or, it looks at independent categories and attempts to speak for the aggregate. Either direction. The whole could look convincing one way, but when variables and categories are considered, the trends change, and vice versa. 

 

An example here is covid-19. 

 

An aggregate view shows that most people survive covid-19 and that the average age for covid deaths is somewhere in the 70s. Age and other health conditions are factors in a person's risk of death. Because of this, the risks and needs for mitigation have often been downplayed. What is not accounted for is that the numbers change when the aggregation is broken down by states, localities, mitigation efforts, and so forth. The aggregate allows generalizations, while the broken down categories show independent severity. For example, the current trends in infection are most prevalent in states and locations where mitigation is most challenged and lax. Dependency on one set of factors can mislead. If we just look at the aggregate for the US, we might feel vindicated to downplay the thing, and if we look at just one of the more severe categories alone, we might overplay things. Context across the board is necessary for a fuller picture or else Simpson's paradox is the fallacy at play.

 

Causal Fallacy

 

There are a few subcategories of causal fallacies, but to each generally, they correlate events or facts to assumed causes. They conflate coincidences or correlate factors to causation. The examples of anecdotal fallacies I mentioned above are also causal fallacies. On one hand, they are observational, but they also attempt to connect apparently related matters together. An example is that gas prices went up most noticeably when Biden took office and he did shut down a pipeline, ergo he is believed the cause of the gas prices. He could have been had the evidence for global covid recovery, supply and demand, and pre-presidency rises in prices not existed as counter evidence. 

 

In short, correlation of seemingly related factors does not equal cause, unless the whole of the data confirms it as such. Until then, it's hyper-assumptive. 

 

Conclusion

 

Much of my angst in online debating is its wading through the mire of misinformation, which is usually related to some sort of logical fallacy. It seems no matter how often I draw from the data, cite sources, reference other possibilities, and so forth, people are often set in their ways. There is often an underlying confirmation bias going on and any correlation of facts that seems to support my opponents' presuppositions is what is held on to tightly. Sadly, I often get accused of being biased myself because I stood against logical fallacies. Sometime I do have to do a double take to ensure that I am treating the info fairly and controlling my own biases, which I am sure I have. Nonetheless, a good debate is where such a challenge causes each party to introspect and consider their own faulty presentations. Yet, when one side continually blockades in their use of obvious or confirmed fallacies, it is hard to have any growth. 

So my advice when debating, and this fall back on me too, is that, "You better check yourself before you wreck yourself." Peace! Out!