djgkanata’s Techdirt Profile

djgkanata

About djgkanata




djgkanata’s Comments comment rss

  • Sep 29th, 2009 @ 3:48pm

    Deeper analysis

    Mike's bottom line is an excellent summary: "...the claims that these measures lead to safer driving aren't supported by the data", because there is no data in the UK report either for or against these claims and indeed the report shows that the data is unreliable.

    However, Mike's claim "injury accidents have actually increased over time" is not supported by the report.

    Part 5 of the report analyzes the reliability of the available injury data and concludes (page 82):
    "The figures presented act as a broad indication of the total number of road casualties in Great Britain, which very roughly illustrates the possible extent to which the STATS19 data are incomplete. However, the limitations of this approximation need to be made clear: ... The nature of these estimates, the way in which they have been produced, the assumptions made and the considerable margin for error all mean that it is not appropriate to produce figures for individual years or to look at trends over time at present, though this may be possible in future."

    In other words -- they have no idea if the number of non-fatal injuries per year has changed over recent years.

    The graph highlighted by the thenewspaper.com article is in Chapter 5 on page 68. Neither the report nor the thenewspaper.com article support the TechDirt article's statement that "injury accidents have actually increased over time". The report does not claim that either line is correct. The report states that different sources of data produce markedly different results, and that it would be unwise to draw conclusions from inaccurate data. The article focuses on the problem of justifying policies by using unreliable data.

    Unfortunately, the bulk of the report is spent presenting data showing year by year trends of many parameters, which is rather undermined by the data reliability analysis of chapter 5.

    Since the year-by-year non-fatal injury trends are unknown, it is not possible to draw conclusions regarding the effect of the traffic control measures on non-fatal injuries. They may be helping, they may have no effect, they may be making matters worse -- there is no data available.

    Part 5 also concludes that the available data on fatalities is accurate and that the number of fatalities per year has decreased over recent years. Anecdotally, the number of traffic cameras has increased over recent years. Correlation does not imply causation. The report does not offer data or opinion on traffic cameras.

    On a separate note, the report does conclude on page 90 "From the in-depth accident studies unbelted vehicle users were found to be significantly over represented when fatalities were investigated...It is estimated that nearly 300 lives would have been saved in 2007 if all car occupants had been wearing a seat belt." The data does support some government-mandated safety measures.