On the Road Again

I know that I should shrug it off as another meaningless study, but the INRIX 2022 Global Traffic Scorecard, as reported in the Philadelphia Inquirer on January 12, 2023, annoys me. Initially it was the article, which accepted this tally of congestion without any analysis of assumptions or methodology. Then I looked up the study on the INRIX website and found that they too gave short shrift to the means by which they reached their startling conclusions. Only by further research could I divine the approach taken, and that is when I really got miffed, realizing that this was another misleading use of statistics.

The Inquirer article told us that Philadelphia was “captive” to the fourth-worst traffic congestion among urban centers in the United States and sits eighth among the world’s cities (OMG, worse than Los Angeles!!!!). Citing the INRIX report, the Inquirer stated that a “typical Philadelphia driver” spent 114 hours stuck in traffic during 2022. That would mean that an average Philadelphia driver sat in their car waiting for traffic to clear for almost 5 days over the last year. Outrageous, if true.

These results “are based on millions of anonymized data points collected from smartphones, GPS systems in cars and trucks, and cities’ own reporting of crashes, incidents, and congestion. Over time, INRIX identified and mapped the most common trip corridors in each urban area…. Tracking travel times on these corridors gives a clear picture of travel times and congestion” according to INRIX. The study goes on to say that the cost of this congestion to travelers is almost $20 per hour for each car.

The questions about this study are legion, though the Inquirer asked none of them. The first is, what is a “typical Philadelphia driver”? Does this only refer to commuters during peak hours? Did they include my 10:30 pm trip to Trenton a to pick up my kids at the train station when there were no traffic delays? Is my wife a typical Philadelphia driver when she’s commuting to Germantown in the morning, but not when she stays late, and heads back into the city around 9:00?

And what is meant by congestion? Do they count every time a car stops for a red light? If you’re travelling the Schuylkill at 45 mph, which would delight any Philly driver, is that congested driving? Is that offset by the rare times when you can zip along at 70? And how do they account for the nut driving down the shoulder at 60 while the rest of us are stopped? 

An article on StreetsBlogUSA did answer some of these questions. Apparently, INRIX segmented their data by time of day and trip characteristics, so presumably we are talking about commuters. Congestion is whenever traffic falls below “free-flow” speeds, which INRIX says it developed using actual traffic data. That would mean a steady 45 on the Schuylkill is congestion. The cost was computed using a $12.81 wage rate, multiplied by 1.13 occupants per vehicle multiplied by 1.37 to reflect the aggravation of sitting in your car, as if aggravation is measurable. If these were the type of assumptions INRIX made, it’s no wonder they don’t highlight them.

More importantly, the Study says nothing about why the congestion occurs. All it does is create an insignificant ranking that generates inconsequential headlines. Actually, that’s probably for the best. The last thing we need are more major road renovations that will take years to complete at incredible cost overruns, slowing traffic even more. The flip side is that if we did that, we could move up to number 6 on the rankings, or even higher (Philly strong)!!!

I know that I have overacted to this drop in the bucket of life. Yet we are inundated with statistics that are similarly suspect, and it is so easy to just accept them as valid. Take something as simple as wind chill factor. Throughout the winter we hear weather prognosticators say, “The temperature is 30, but it’s going to feel like 10 because of the wind.” NO, IT’S NOT. Wind is not a constant. It gusts. It may feel like 10 or below when you’re walking through a center city wind tunnel, but it won’t feel that way if the gusts calm, and you’re walking in the sun.

The pandemic was prime time for statistics. We would routinely hear that COVID rates had doubled in an area over the last week but were never sure if that meant that they went from 2 to 4, or 1000 to 2000. It often depended on who was promulgating the statistics, and what they wanted to accomplish. Sifting through the mass of data is impossible, which means you have to rely on the good faith of those that do, and that can be problematic, to say the least.

The malleability of statistics is a big reason why so many discussions on crucial topics like climate change come to a scratching halt. Statistics about carbon dioxide in the atmosphere, or the rise in sea levels are thrown around like nerf balls, bouncing harmlessly off readers’ noggins. Whatever those statistics say, someone will conflate them into an immanent world collapse, or deflate them into a meaningless blip.

The problem is that it takes overstatements to get most people’s attention. A nuanced approach may work well on some late-night talk show (back when we had such shows), but if you want newspaper headlines, or shared Facebook posts, you better be controversial and extreme. In the end all this does is provide fodder for dissension, not consensus.

So, what can we do? One thing is to discern when statistics help explain something and when they don’t. Some trumped up congestion rate adds nothing to our understanding of traffic flow. We know that traffic can be bad during rush hour, and at any other time if you are unlucky enough to hit an accident or road construction. Putting a number on it is meaningless. Just like you know that you better bundle up if the wind is blowing in the middle of January without a wind chill factor.

Also, look at the source. When the CDC gives you trends during a pandemic, best to pay attention. When it’s a friend from high school passing along climate change information issued by the American Petroleum Institute, be skeptical. Sometimes that takes some searching, because the API is probably issuing those statistics through Americans for a Cleaner Tomorrow, but it’s worth the effort.

Finally, look at methodology. If those issuing the statistics make you look hard for their assumptions, like INRIX, they probably are not a good source. Those above board will proudly highlight how they reached their conclusions and provide guidance on what you can do in response.

I know I jumped on my soapbox here, but the flood of worthless stats we see is one of the those burrs in my side that I just can’t shake. It’s just so easy to be misled. I know that I have, many times. Oh, the glories of the modern world.