Photo: Flickr user resiak

An essay by philosopher Harry Frankfurt, first published in the mid-1980s but re-released in book form in 2005, offered what was perhaps the first serious analysis of a topic near and dear to any political observer’s heart: bullshit. Frankfurt argues that bullshit differs from lies in one important way. The liar believes that there IS such a thing as the truth: it’s the thing that the lie is designed to conceal. In contrast, a bullshitter has no regard for the very idea of truth. To a true bullshit artist, words are little more than a tool to achieve some arbitrary purpose — to seem informed, for example, or to manipulate other people so that they’ll do what you want them to. Whether words are true or false is beside the point; all that matters to the bullshitter is the effect that his or her words achieve.

Frankfurt’s essay came to mind when I read a WSDOT spokesperson’s recent article about traffic growth on the SR-520 bridge across Lake Washington. The spokesperson was responding, indirectly, to one of my previous posts pointing out that WSDOT data shows zero traffic growth across the SR 520 bridge for the last 15 years.  Here’s his rebuttal:

[Sightline’s] analysis didn’t capture a complete picture of traffic on the corridor. Its analysis starts in 1996, the peak year of growth and traffic over the last 25 years, and continued through 2010, capturing the dot-com bust between 2000 and 2003 and the subsequent national recession, both with significant traffic effects in this region.

In contrast, the Washington State Department of Transportation’s (WSDOT) longer view of traffic data on the SR 520 bridge begins in 1984 and projects to the year 2030, capturing nearly 30 years of historical data to produce a more accurate trend. Traffic generally rose annually since 1984, until the beginning of economic uncertainty in the middle of the 2000s.

WSDOT’s response is about as pure an example of B.S. as one is likely to find in public discourse. Let’s break it down, point by point.

  • Give today to help Sightline reach our goal of $100,000!

    Thanks to Jennifer Snyder for supporting a sustainable Cascadia.


    $77,000

    Donate Today
    • [Sightline’s] “analysis didn’t capture a complete picture of traffic on the corridor”. B.S. I have, in fact, looked at traffic trends in the complete cross-Lake Washington corridor, including both SR 520 and I-90. As far as I can tell, the corridor-wide trends tell essentially the same story as the SR-520-only trends: a long-term flat-lining of traffic across Lake Washington. For I-90, the flat-lining started in about 1998.
    • “Its analysis starts in 1996.” B.S. I did, in fact, consider the traffic trends prior to 1996. They show rapid growth through the mid-1980s, slow and intermittent growth from the mid-1980s to the mid-1990s, and then a flattening out of traffic (or possibly a slight decline) starting in the mid-to-late 1990s. Here’s a long-term chart of daily traffic, showing two different WSDOT data series, one for weekdays only and another with averages for all days of the week. (But please see my caveats about the numbers at the end of this post!) 
    • “1996, the peak year of growth and traffic over the last 25 years.” B.S. For both WSDOT data series, the nominal peak in SR-520 cross-lake traffic occurred in 1998, not 1996 — a fact that would have been easy to ascertain if the WSDOT spokesperson had actually looked through his own agency’s data. But regardless of whether the nominal peak occurred in 1996 or 1998, the data show essentially the same thing: traffic hasn’t budged in well over a decade.

    Data from US Bureau of Labor Statistics

    • “capturing the dot-com bust between 2000 and 2003 and the subsequent national recession, both with significant traffic effects in this region.”  That’s bullshit — not because it’s false, but because it’s selective and misleading. The years of flat traffic on SR-520 include not only busts (2000-2002, 2008-2010) but also booms, including the hot dot-economy of the late 1990s and the years of steadily rising employment from about 2003 through mid-2008. (See the chart of metro-Seattle employment trends to the right.) By attributing the flat-lining of traffic to economic busts, WSDOT implies that traffic will grow again once the economy picks up. But that ignores the fact that traffic remained flat through extended economic booms in the 1990s and 2000s. That sort of selective use of evidence is classic B.S.
    • “In contrast, the Washington State Department of Transportation’s (WSDOT) longer view of traffic data on the SR 520 bridge begins in 1984” B.S. To me, this sentence implies that WSDOT traffic forecasters looked at traffic trends on SR-520 from 1984 through 2010, and used those traffic trends to extrapolate future growth. But that’s not how those projections are made. WSDOT (or its consultants) use a traffic model for all of Puget Sound that extrapolates traffic volumes in general, based on projections for employment, population, and income growth. The model then assigns trips to various destinations, modes, and roadways, and is customized to give a finer-grained look at the 520 project area. (For more details, see chapters 3 and 4 of the transportation section of the SR-520 Final EIS.) But the regional traffic model has done a notably poor job of predicting recent traffic trends, both in general and on specific roads. In particular, the model tends to over-assign traffic to roads that have actually reached saturation. Regardless, the SR 520 projections aren’t based on a “longer view of traffic data on the SR 520 bridge,” but on a much more complicated model that assigns traffic to SR 520 in ways that can essentially ignore real-world traffic trends on the bridge.
    • “capturing nearly 30 years of historical data to produce a more accurate trend.” There is no evidence that “capturing nearly 30 years of historical data” actually helped WSDOT produce “a more accurate trend.” Indeed, as the chart in the original post (below) shows, WSDOT’s traffic projection methods have resulted in a series of forecasting fiascoes. As I’ve said before, the 1996 projection was completely forgivable; and you can even excuse the 2002 projections, since memories of rapid traffic growth were still fresh. But by 2011, it has become abundantly clear that WSDOT’s traffic projections aren’t passing the laugh test. Besides, claiming that a trend is “more accurate” before it’s been confirmed by actual events is B.S. of a particularly noxious sort.

    • “Traffic generally rose annually since 1984, until…the middle of the 2000s.” Ungrammatical, but still B.S. WSDOT’s Annual Traffic Report found year-over-year traffic declines in 1993, 1996, 2000, 2001, 2002, 2004, 2005, 2006, 2007, and 2008. WSDOT’s Ramp and Roadway Report is only released every other year, but found biennial declines in 1990, 1992, 2000, 2002, and 2008. Similarly, the Seattle DOT reports data only for 1996 through 2004, but found year-over-year traffic declines in 1997, 1999, 2000, 2002, and 2004. In short, traffic generally rose annually until the middle of the 1990s, after which came a mixture of slight increases and slight decreases, depending on which year and which data source you look at.
    • “the beginning of economic uncertainty in the middle of the 2000s.” Gaah! Greater Seattle’s economy was humming along nicely in the middle of the 2000s, with employment rising and unemployment rates occasionally dipping under 4 percent. Employment didn’t start to fall in earnest until late 2008. (See the employment chart, above.) It’s like WSDOT isn’t even trying to get the facts right!

    Frankfurt’s essay was intended as a warning. He believed that B.S. does more damage to public discourse than lying, because bullshit undermines the very idea of truth as a useful concept. But if anything, WSDOT has ignored the cautionary notes in Frankfurt’s essay, and instead has decided to interpret “On Bullshit” as an instruction manual.

     

    (Wonk alert: be careful with data from 1988 through 1990. From what I can tell from the archived pdfs, that’s when WSDOT transitioned from measuring traffic counts by counting axles to counting vehicles — so the numbers in those years are a bit screwy. I have chosen to omit two data points from the Annual Traffic Report series during that period, because they seem to be anomalies and WSDOT stopped reporting data for one of those years in subsequent editions of the report. In short, there’s reason to view traffic trends in the late 1980s and early 1990s with some skepticism, and the numbers before and after 1990 may not be directly comparable.)