The primary source of automobile crash statistics in any state are the reports filed by the investigating officer. In Texas, this form is known as the Texas Peace Officer’s Accident Report or an “ST-3” (or ST-3C in the case of a commercial vehicle supplement). Among the information collected on this form are the identities of the people involved, vehicle information, the location of the crash, and the road conditions. In addition there is a narrative description and diagram of the crash that presents the officer’s opinion of what happened. As described in the Instructions to Police for Reporting Accidents (1996), the narrative:
When coupled with the diagram, …must tell the main events of the accident, describe the major factors and, so far as possible, tell how it happened (emphasis added).
It is the officer’s diagram and description of what happened that gives the most useful information in reconstructing the events.
On the ST-3 two roads are identified, the road on which the crash occurred and the intersecting street. The second street is only identified if the crash occurred in an intersection or at a railroad crossing (Instruction, p.2). For each of these streets there are two boxes (one for “yes” one for “no”) with which the officer indicates whether the crash occurred in a construction zone(1). If both streets, the one on which the crash occurred and the intersecting street were under construction (that is, if the crash occurred in an intersection that was under construction) then both boxes would be checked “yes”.
When the data submitted on the ST-3 is entered into the data files at the Department of Public Safety (DPS) in Austin, Texas, the information provided by the “check boxes” as well as the information provided in the narrative are used to determine how the information is coded.
There are four codes that can be entered by DPS that indicate a construction zone. These are:
- In highway construction area – Not construction related.
- In highway construction area – Construction related.
- In other construction area – Not construction related.
- In other construction area – Construction related.
In this context “other” construction areas are non-highway construction events such as work along the side of the road to replace a utility pole.
The distinction between a crash that is coded as “construction related” and one that is coded as “not construction related” is critical in any discussion of work zone crashes. The following tables illustrate the vast differences in the magnitudes of the numbers associated with these two concepts. Table 1 – Crashes Occurring in Texas Work Zones presents data on crashes that occurred in work zones. This table was derived by combining the data for all four of the categories show above. This is virtually the equivalent of all the ST-3 forms for which a “yes” box was checked for either of the two roads.(2) This table indicates (for example) that in 1996 there were 107 fatal crashes that occurred in work zones. While not depicted in the table, these 107 crashes resulted in 123 deaths. This number, 123 is different from the 111 deaths reported by the Fatal Analysis Reporting System (FARS) maintained by the National Transportation Safety Administration (NHTSA) for the same time period.
It has been suggested that the differences noted are due to different coding rules or different “cut-off” dates for the two different system. Regardless of which of these numbers is used, Texas comes out on the top of any list that ranks the states by deaths occurring in work zones if either of these numbers is used. However, there is a real question as to whether it is appropriate to use numbers based on a variable that simply indicates that a crash occurred in a work zone. A hypothetical example will be used to illustrate the problem.
Suppose another state does not have the previously discussed “check boxes.” In this state the coders at the equivalent DPS facility use only the narrative to distinguish the role of the roadway in the crash. This is the equivalent of selecting those records from the Texas files that are coded as “construction related” because regardless of whether a box is checked “yes” only those crashes for which the narrative indicates the work zone contributed to the cause or severity of the crash are coded as construction related.
Table 2 – Texas Work Zone Related Crashes presents data on work zone related crashes. This table was derived by combining the data for the two categories of construction related crashes. This table indicates that in 1996 there were 12 fatal crashes in work zones. And, again while not depicted, these 12 crashes resulted in 18 deaths. While in no way minimizing the importance of even one death, the 18 deaths resulting from this analysis are a far cry from the 111 reported in the Fatal Analysis Reporting System.
Though it is admittedly difficult to obtain comparable information for different states, at a minimum, when discussing work zone crashes, more information regarding how the data are collected, and how the determination of the involvement of the work zone has been made need to be taken into consideration.
1 Throughout this discussion the terms “work zone” and “construction zone” will be used interchangeably.
2 It is also possible for a crash to be coded as occurring in a work zone if the narrative mentions a construction zone, but the box had not been checked. This, however, is an extremely rare event. Perhaps even more unlikely is the possibility that one of the “yes” boxes was checked and the crash was coded as not having happened in a work zone. In all likelihood the latter would be an example of a coding error.
The information presented here was derived from data provided to the Texas Transportation Institute by the Texas Department of Public Safety and the Texas Department of Transportation.
Last verified: 7/20/2007