June 16th, 2009

Around a week ago, a local news item appeared on the website of NBC-DFW publicizing the traits of “Safe Passing” legislation awaiting the Texas Governor’s signature. Within this blurb, a statement was made suggesting this Bill would “curb the increasing number of bicycle injuries on North Texas streets.” This seemed an odd comment given the fact that, to my knowledge, no recently publicized reports or studies had come to this conclusion. I criticized this and other aspects of the post and, given the apparent authority in the context, I decided to do a little investigation.

Much to my surprise, I discovered there is no readily available source for information regarding injury rates. An inquiry to the county injury prevention center at the Dallas County Hospital District revealed there had never been a request to produce a summary such as this. That perplexed me, since, in order to come to the conclusion reached by the NBC-DFW reporter, Holly Lafon, one would have to request this information. Being the ever curious scientist, I initiated a request for the data required to derive these numbers and will post the results, should they be forthcoming, at a later date.

In the meantime, I thought I would take a look at a somewhat more critical data set: fatality rates.

The primary repository is the Fatality Analysis Reporting System (FARS) Encyclopedia of the NHTSA. FARS is a freely available resource containing data on all vehicle crashes that occur on a public roadway and involve a fatality in the US. Though far from perfect in its accounting methodology, this database is a useful tool for gaining insight into the numbers of vehicle deaths.


Data for this analysis is derived from the FARS database. For the purposes of this overview, no attempt was made to carve out only those numbers relevant only to adult cyclists riding a bicycle for utility or transportation. Therefore, both children and fatalities falling under the somewhat ambiguous designation of “Other Cyclists” are also included. Since this practice was preserved across all surveyed data sets, not refining the focus should have little bearing on the overall point of the discussion.

The primary focus of research was the greater DFW metropolitan region. Since the majority of transportation cyclists live in either Dallas or Tarrant County, surveys were restricted to those geographic data subsets, rather than all of North Texas – which, according to NCTCOG, comprises a sixteen county region.

Much is made of the safety benefits of bike lanes and other protectionist facilities. Accordingly, surveys were also made of three other large cities in Texas: Austin, Houston and San Antonio. Austin and Houston are often cited as exemplary for their encouragement of bicycle commuting through installation of bike lanes. Though more fragmented and sometimes criticized, San Antonio also makes use of the s facilities to promote safe cycling.

Finally, fatality and injury figures were surveyed for Portland, Oregon. Though differing greatly in climate, population, population density and cultural philosophy, this city is often held up as the goal to which all other cities should aspire.


The FARS data, at the time of this survey, spans the years 1994 through 2007. Fatality totals were mined for the entire state of Texas and each of the target municipalities during each of these fourteen years. The statewide total was used as the base by which to compare the numbers for each of the cities analyzed.

fatalities, Dallas

fatalities, Dallas

Though the total number of fatalities in the state fluctuates widely, traffic related deaths in Dallas County are relatively stable. With the exceptions of 1997 and 2004, rates hover between one and three fatal interactions per year.

fatality rates, DFW

fatality rates, DFW

Again, with few exceptions, the numbers of fatalities are just about equal between both Dallas and Fort Worth. A couple of anomalies exist in 1995 and 1997, when Fort Worth and Dallas, respectively, had higher than average numbers of deaths.

Despite the assertions by some Fort Worth bicycle advocates, Dallas’ neighbor to the west affords little advantage in terms of safety to transportation cyclists.

fatality rates in major Texas cities

fatality rates in major Texas cities

Interesting results begin to appear when Austin, Houston and San Antonio are added to the mix. Most apparent is the high volume of deaths in Houston as compared to the other cities. With the exception of 2004 and 2006, Houston weighs in with fatality numbers equal to or exceeding all other surveyed municipalities combined.

fatality rate comparison

fatality rate comparison

Despite numerous arguments from cyclists from the state capitol, the numbers do not lie: Dallas and Austin have almost identical fatality rates. Exceptions oscillate back and forth. In 2001 and 2002, Austin had significantly higher rates of cyclist deaths than Dallas. Conversely, 1997, 2003 and 2004 were bad years for North Texas. Notable is the lack of any reported fatalities for Austin in ether 1997 or 2003.

Twelve years of bike lanes in Austin has apparently made little difference when it comes to fatality rates.

fatality rate comparison

fatality rate comparison

Facilities proponents like to point to Portland as a bicycling nirvana. Their liberal use of segregated facilities and experimental enhancements certainly keep them in the limelight. Though they flaunt a high bicycle ride share rate, which they attribute to their infrastructure, this has done little to lower the rate of fatalities.

Apologists like to highlight the decreasing proportion of fatalities to overall ridership numbers. However, they fail to produce anything other than anecdotal evidence to support their totals and the fatalities speak for themselves. Germane to this discussion is the fact that average fatality rates for Dallas and Portland are almost equal at 3.00 and 3.07, respectively.

Note: The totals for 2008 are unverified. Bike Portland has claimed there were zero fatalities for 2008. Meanwhile, the six deaths indicated for Dallas derive from published news items appearing in the Dallas Morning News. Confirmation will require awaiting the release of 2008 FARS data later in the year.


In terms of fatality rates, the data presented in this survey has proven there is no trend, up or down, in any of the municipal areas surveyed. Averaging the numbers from all years reveals, with the exception of Houston, there is very little difference in bicycle transportation related deaths within any of the cities. This is important to note, since many facilities advocates suggest that segregated bike lanes enhance safety. These data prove otherwise.


A few will argue that this information compares apples to oranges in terms of refuting the NBC-DFW piece. That author cited increasing injury rates, while I compared fatality rates. One could argue that the one is reflective of the other, if, all things being equal, fatality rates are a consistent subset of overall rates of injury. However, in the absence of empirical evidence to corroborate that belief, it would be disingenuous to draw such a conclusion.

By focusing on fatality rates, this survey has succeeded in refuting one aspect of the safety argument. If one assumes mortality to be an equal gauge of safety as injury rates, then the presence or absence of bike lanes has little effect on survivability. Annual death rates attributable to bicycle/motor vehicle collisions are relatively consistent within a given region through time. Though one can certainly appreciate occasional fluctuations, the overall rates remain the same.

Facilities proponents need to understand that it does nothing to further their cause when resorting to the use of hyperbole and hearsay to justify their beliefs. If the arguments in favor of bike lanes and paternalistic legislation must rely upon innuendo then there is a flaw in the premise.

revision log:
20090619 – added a paragraph to the section on Methodology better explaining the scope of the Person Types surveyed and the reasoning behind the decision.

3 Responses to “Carnage”

  1. pmsummer says:

    Very good research, Herman. What’s missing for analysis is demographic/collision type, as there is a dramatic difference in type between adult and juvenile bicycle fatalities (not news to you, I know).

    I trust KXAS will be getting back to you on this.

  2. Herman says:

    Thanks for the feedback, PM.

    I agree, in principle, with your criticism. However, I opted not to slice that differential out at present for two reasons. The first was in the interests of time. I felt it worthy to get this partial rebuttal out in as timely a manner as possible. As I alluded to in the post, parsing the FARS data set is less than ideal. It is impossible to rerun the same query conveniently for different years without fully selecting the filtering criteria for each iteration. Unless I am missing something, there is also no means of generating a report which includes the sub-totals for each sub-query within each annual tabulation. Since I was looking at seven counties, spread across two states and Texas as a whole, you can, I hope, appreciate and forgive the oversight this go ’round.

    The other reason is more cynical. You know as well as I that the media and the CIC, facilities whiners are not going to make that distinction. They are going to lump every conceivable demographic into the results in order to prove their point. I will need to revise the narrative to reflect this, but, in a way, I did the same thing. The FARS data set includes two Person Types applicable to this discussion, “Bicyclist” and “Other Cyclist”. Both were included in my analysis. I currently have no idea what an “Other Cyclist” is, but those are among the results – applicable or not.

    With the exception of possible skewing with respect to the Harris County numbers, I think one can appreciate the numbers of “cyclists” killed annually – all inclusive of experienced vehicular cyclists, uneducated cyclists, children and “other” cyclists – is so low as to be almost insignificant. Having gotten the general picture released to the wilds, I will delve deeper into the results and perform a more critical analysis. More people died from simply falling than were killed by a motorist while riding their bike in 2006. Nothing is 100% safe and a little perspective is needed.

  3. Steve A says:

    Wouldn’t a simpler test of facilities be to simply plot facilities spending per capita against cycling fatalities per capita? Such a plot shows no reduction in fatalities as expenditures rise. I have not, however, attempted to drill down to local cases in such a correlation.

RSS feed for comments on this post. And trackBack URL.

Leave a Reply

You must be logged in to post a comment.