The Target was painted around the Arrow

With our world becoming more digital, more automated, more directed by advanced analytics, deep learning, and artificial intelligence, it brings up a question in my mind about just how good are we at using data? Many situations from crime statistics, to environmental (air and water quality) data to economics (inflation, employment, etc.) are ending up in government regulations and often lawsuits. How well do politicians, lawyers and even average citizens understand how data (we now have a lot from a number of different sources) is used to resolve different opinions and settle compliance complaints.

According to a survey by the Nuffield Foundation, a British think-tank, in England fewer than one in five students studied math after 16. In 18 (of the 24 OECD countries surveyed) more than half did. In eight of these countries, everyone did. UK government data suggest that almost half of the working-age population in Britain have the numeracy skills of a primary school child.

How about the numeracy skills of our leaders? Statistics on STEM qualification for government employees are not very well known (or measured). But most estimates put them at about 5% in the UK, about 16% in the US but in South Korea the level is about 30%. Lawyers are little better. According to Dame Kate Bingham, who chaired Britain’s vaccine taskforce, how can a modern society run a business development in the 21st century with “nobody who know anything about business or science?”

In a lecture in 1959 titled “Two Cultures”, scientist and writer C.P. Snow warned that society was “being split up into two polar groups. Those who understood science and those who did not.” It is even more serious when people in power don’t even know what they don’t know. That observation is even more true today than it was 65 years ago.

Daniel Patrick Moynihan, who died in 2003, famously said that “people are entitled to their own opinions but not to their own facts”. He also liked to quote the maxim that “it’s not ignorance that hurts so much as knowing all those things that ain’t so.” In today’s social media dominated world, we can get our own facts that support our specific opinions and even defend those “truths” with statistics.

According to Wikipedia (https://en.wikipedia.org/wiki/Texas_sharpshooter_fallacy ) the phrase of “painting the target around the arrow” is sometime referred to as “the Texas sharpshooter fallacy is an informal fallacy which is committed when differences in data are ignored, but similarities are overemphasized. From this reasoning, a false conclusion is inferred. This fallacy is the philosophical or rhetorical application of the multiple comparisons problem (in statistics) and apophenia (in cognitive psychology). It is related to the clustering illusion, which is the tendency in human cognition to interpret patterns where none actually exist. The name comes from a metaphor about a person from Texas who fires a gun at the side of a barn, then paints a shooting target centered on the tightest cluster of shots and claims to be a sharpshooter.”

With all the data available from so many sources these days (from science but also from social media), a lot of us are claiming to be Texas sharpshooters, or should I say data experts, but few of us have taken the time to get a PhD in Statistics or Applied Mathematics. It is getting easier to learn how to develop an artificial neural network model in a programming language like Python, but do we really understand what the model is trying to tell us? Today the technology is large language models, and generative AI. Most powerful tools but do we know what we are aiming at before we train the model?

Again, from that same Wikipedia article: “The Texas sharpshooter fallacy often arises when a person has a large amount of data at their disposal but only focuses on a small subset of that data. Some factor other than the one attributed may give all the elements in that subset some kind of common property (or pair of common properties, when arguing for correlation). If the person attempts to account for the likelihood of finding some subset in the large data with some common property by a factor other than its actual cause, then that person is likely committing a Texas sharpshooter fallacy.”

Most people are infatuated by the new programming models, the GPT algorithms that “eat the internet” and sometimes other people’s intellectual property. Thank you OpenAI for this gift. The technology is the shiny new object but do we really understand the problem? Or is it “shoot then aim” rather than “aim then shoot” at data-centric problems. My recommendation is to watch out for Texas sharpshooters, check the data and understand what problem you are trying to solve.

(thanks to an article in the August 24th edition of the Economist and an article titled “Two Cultures” for this topic)