The Data of Accident Law – What It Means for You

Welcome back! In this series of blogs, we look at technology trends and data-driven analyses impacting the legal landscape of automotive and other accidents in Texas. These blogs are intended to bring the story of historical and current trends to life, with specific emphasis on what these trends mean for you and, most importantly, to remind you of the kind of customized help Tate Accident Law and Tate Law Offices provide when you or a loved one is involved in an accident. 

Understanding the Scope of Traffic Accident Data

Let’s start off with a not-so-fun fact. When I use statistics to talk about traffic deaths in Texas, I’m primarily relying on data from the Texas Department of Transportation. Their data only includes accidents on public roads. An accident that occurs on a private road, in a parking lot, in a driveway, or elsewhere outside their purview is not accounted for in their data. So when I talk statistics I’m sadly under-representing the true number of accidents and fatalities. Why does the distinction matter? Because no matter where an accident occurs, whether on public roads or not, you still need to take the same precautions. On our website you can find some great information on exactly how Tate Law can help you in the aftermath of an accident. But don’t wait until you’ve had an accident to learn what to do in an accident. In another not-so-fun fact I’m repeating from a previous blog post, it is extremely likely that you or someone you know will be involved in a traffic accident. As in inevitable. Educate yourself, but Tate Law in your phone contacts? Not a bad idea. 

In the last blog, I talked about the number of Texas traffic fatalities per year, and made the following point: though the number of traffic fatalities has increased over time, by one key measure the relative number of deaths has improved. Upon reviewing some traffic statistics in preparation for this blog, a new story emerged. The new story casts the last blog in a different light and takes us down a totally different path than I had initially intended for this discussion. But the path is an interesting one, and the story deserves to be told. 

Warning: bad math humor ahead. There’s an old saying that statistics don’t lie, but statisticians do. There’s another old saying that statistics don’t lie, but liars use statistics. If you re-read my last post and compare it against this one, you might think two different people wrote them. Is one a liar? Not necessarily. At least I hope not, because if either is a liar then I’m the liar. But the lesson here is that the same numbers can be interpreted differently, and even statistical analyses and data science have a layer of subjectivity. 

The Psychology Behind Data Interpretation

I promise I’ll get to the real story soon, but humor me for a minute on the psychology of statistics. Yup, I said psychology of statistics. Let’s assume for the sake of discussion that I’m an honest person and I’m not trying to lie my way through data analyses, cherry-pick numbers or otherwise deceive. Even so, since I’m not a robot I can bring subconscious biases into a data analysis. I could, for example, unintentionally look at a set of data that’s not fully representative of the population of interest. We call this selection bias. I could have my own expectations of what conclusions will come of an analysis before starting the analysis, and conduct my research in a way that reinforces what I expect to be true and shuts down the part of my brain that should be completely open to alternatives. We call this confirmation bias. There are lots of other forms of bias that humans can demonstrate, and if you ever want to read an incredible book on the topic, I recommend Thinking, Fast and Slow by Daniel Kahneman. 

Now, there’s another side to being a data scientist that may surprise you. A good data scientist is like a good lawyer. What I mean by that is: the best lawyers can effectively argue either side of a case. Presented with the same facts, a great lawyer can construct different logic trails and different frameworks to draw different conclusions depending upon which side of the case he/she is on. (Just so you don’t misinterpret my analogy here or read between the lines in the wrong way, let me say clearly about my friends who are practically family at Tate Law that they consistently represent the victims of DUI and other accidents.) So, to close out my analogy here, a good data scientist can construct different interpretations of the same data to provide different contexts and conclusions. 

So where am I going with all the talk about the psychology of statistics and data scientist and lawyer analogies? I wanted to give some insight into how the upcoming analysis can use the same data as my last blog and yet seemingly arrive at different conclusions. Are these conclusions different because I’m a great lawyer or because of unconscious biases? Let me put it this way: I’d make a terrible lawyer. But, this analysis isn’t really me saying the previous analysis was wrong. It’s more about looking at the same data through a different lens, applying the same data science principles and logic, and recognizing that two seemingly opposing truths can co-exist. 

Same Data, Different Story: A Closer Look at the Numbers

Let’s do some data analysis, shall we? Here’s what we’ll do this time around. I’m going to show you the same numbers as last time and revisit some of my previous analysis. The below data from the Texas Department of Transportation shows that the number of deaths in 2023 (4,283) increased from 2003 (3,822), by a total percentage of 12%. However, the deaths per 100 million miles traveled metric shows a decrease from 2003 (1.75) to 2023 (1.45), a percentage decrease of 17%. 

Based on this data I previously said: “By this measure of deaths per miles driven, driving in Texas has become relatively safer in the past two decades, at least from a death toll perspective.” I stand by that analysis. But now let’s look at the data differently. 

First, let’s play a little visual trick. In the below chart, which represents the exact same data as the above chart, I’ve changed the vertical axis to a different range – effectively I’m just zooming in on the data. 

Kind of a different tale, right? It’s still true that we see an overall decrease from 2003 to 2023, but that’s not really what jumps out from this chart. There’s a major downward trend from 2003 through 2011, then a spike in 2012 that tapers down through 2019, then another spike in 2020. Now let’s play another visual trick. This one’s less of a trick – let’s just focus on 2019 through 2023.

So in my previous analysis I said: “By this measure of deaths per miles driven, driving in Texas has become relatively safer in the past two decades, at least from a death toll perspective.” What I said is completely true. But when we look at the above chart, the story is one of a dramatic increase in deaths and decrease in safety over the past 5 years. Why? What’s the story there? Hint: remember that year when we all started wearing masks out in public? Yup, you got it – Coronavirus 2019, also known as COVID-19.

COVID-19 and the Unexpected Rise in Traffic Fatalities

An obviously tragic thing happened when COVID hit. While we’re on the topic of liars and statistics, let’s talk about what would seem a simple task – counting the number of COVID deaths. If you read different sources for statistics on COVID deaths in the U.S., you’ll find that numbers vary. One seemingly simple and obvious measure is counting deaths based on the number of death certificates listing COVID as cause of death. However, many who died of the disease weren’t tested, so the death certificate number must underestimate the total count. RIght? Maybe. COVID may coexist with other potentially lethal diseases, so death with COVID doesn’t necessarily by COVID. By that logic, maybe some number of deaths attributed to COVID overestimated the count. 

But what does COVID have to do with fatal auto accidents? First let’s recognize the old statistical adage that correlation is not causation. We see a spike in COVID deaths and a spike in traffic deaths at the same time, but we can’t say that COVID caused car accidents or car accidents caused COVID. More bad math humor ahead. In the summertime as the weather warms, ice cream sales increase. Shark attacks also increase. Eating ice cream does not cause shark attacks.

But why the drastic increase in tragic automotive deaths at the same time as COVID deaths? Here again the answer isn’t simple. Remember how empty the streets were at the beginning of the pandemic? I do. At that time, I was living in Connecticut and commuting to Manhattan. Here’s what Times Square looked like at the onset of the pandemic. 

Here’s what Times Square usually looks like.

Thankfully I’m no longer working in the city jungle and I’m living back home in Texas again, but scenes were similar here. For a time, commuter traffic dipped substantially. Notice I intentionally said commuter traffic. Trucking traffic didn’t dip at all. I mean, we weren’t going to live without Amazon deliveries, right? Especially when we were stuck at home with nothing to do but eat and binge-watch Netflix. 

According to the Texas A&M Transportation Research Institute, commuter traffic dipped by 50% for the initial few months of the pandemic, but rebounded in the months thereafter. So why did automotive deaths increase? Several reasons. Emptier roads in the beginning of the pandemic led to more speeding, and more fatal accidents. Increased anxiety during the pandemic led to more alcohol and drug use, and translated to more DUI-related deaths. The Federal Motor Carrier Safety Administration (FMCSA) relaxed restrictions on the number of hours truckers could drive, leading to increased fatigue and more trucking deaths. 

So let’s apply a little statistical intuition. If traffic deaths increased while overall traffic decreased, we’d expect some measure to reflect that the increase in traffic deaths was somehow disproportionate. Indeed we can look at a couple numbers and see exactly that. Let’s just look at 2019 vs. 2020 (the pandemic started in March of 2020).

*DPVMT = Deaths Per 100,000,000 Vehicle Miles Traveled

Deaths increased by 7.6%. DPVMT increased by a disproportionate 19%, a reflection of the fact that the increase in deaths was at the same time as a decrease in overall traffic.

So, What Does This Mean for You?

As always, I want to close out the data story with an answer to: what does this data mean for you? And usually the answer will be the same: know Tate Accident Law and Tate Law Offices are ready to serve you and anyone you know who has been in an accident. And you should know Tate Law is ready to serve anybody you know who may be in an accident – which statistically is everyone!  

About the author: The author works exclusively with Tate Accident Law and Tate Law Offices to explore technology and data trends in the context of accident law. Trained as an engineer, he holds Bachelor of Science degrees in Bioelectrical Engineering, Applied Mathematics and Computer Science from Southern Methodist University with summa cum laude honors. He followed his undergraduate studies with Master’s and Doctorate degrees from the Massachusetts Institute of Technology, where he was named a Presidential Fellow. Upon completing his graduate work in the field of real-time data analysis systems and Artificial Intelligence at MIT, he has spent the past 25 years designing data analytics platforms and teaching technology and data subjects to beginners and experts alike.  

Geographies served by Tate Accident Law: Grayson County, Denton, Whitesboro, Sherman, Plano, Mesquite, Anna, McKinney, Gainesville, Dallas, Durant, Bonham, Denison, Garland, Richardson, Allen, Celina, Prosper, Irving, Fort Worth, Ardmore, The Colony, Garland, Richardson, Lewisville, Little Elm, Anna, Melissa, Parker, Addison, Frisco, Plano, Wylie, Mesquite, Allen, McKinney, Murphy, Prosper, Irving, Rockwall, Grapevine, Arlington, Carrollton, Grand Prairie, Melissa, Addison, Cleburne, Bedford, Hurst, North Richland Hills, Watuga, Burleson, Haslet, Mansfield, Benbrook, Haltom City, Kennedale, Saginaw, White Settlement, Crowley, Westlake, Azle