General Instructions and Helpful Information

Regarding the “Expanded” List of Variables

Trucks Involved in Fatal Accidents (TIFA) Dataset, 1991-2004

Used in Appendix V Analysis of the 2007 Hours of Service Regulatory
Impact Analysis

January 23, 2008

The accompanying Excel spreadsheet, titled “Expanded List of
Variables_TIFA_1991-2004,” contains several additional variables to
the “basic” Trucks Involved in Fatal Accidents (TIFA) datasets
representing calendar years 1991-2005 that were submitted in December
2007 to the Hours of Service (HOS) docket.  This raw data is being
placed in the docket in response to a telephone request by a member of
the public received on January 22, 2008.  It includes variables that
were used in the supplemental analysis conducted and documented in
Appendix V of the 2005 and 2007 HOS regulatory impact analysis (RIA),
which is currently in the docket.  The basic TIFA data sets submitted to
the docket included variables for “truck driver fatigue” and
“hours of driving at time of crash” and were used in analyses
conducted and documented in the RIA.   

Below is general information on the variables included in this dataset,
which contains information on large trucks involved in fatal crashes
between the years 1991 and 2004.  While general information on the
variables included and analysis of this data are contained below, for
technical assistance, users are encouraged to contact Dan Blower,
Director of the Center for National Truck and Bus Statistics, University
of Michigan Transportation Research Institute at (734) 763-6079. 

1) There is no variable to record weather as a contributing factor--at
least not systematically. The variable for weather is included in this
data set, but simply records the ambient conditions at the time of the
crash.

2) The fatigue and distraction variables in this dataset are recoded
from the Fatality Analysis Reporting System (FARS), using the driver
contributing factors variables. 1 is yes, 0 is no.

3) Mechanical condition is recoded from the FARS vehicle factors
variables. If *any* mechanical condition was coded by FARS, I set the
variable to 1, otherwise zero.

4) Note that in the “basic” TIFA data sets contained in the HOS
docket (years 1991-2002, and 2003-2005), cases were excluded where the
vehicle was government owned or where the vehicle was a daily rental and
GVWR was less than class 7, since they are not generally subject to
FMCSA regulations.  In this dataset, “all” cases are included, but
we have also enclosed the variables to do the exclusion, if that is
desired.

e.g., if v1058=6 then delete

       if v1058=7 and v1050<7 then delete.

 

5) The code labels for weather are included below; note that this simply
records the condition at the time of the crash, not contribution.

     1='Normal'

     2='Rain'

     3='Sleet'

     4='Snow'

     5='Fog'

     6='Rain & Fog'

     7='Sleet & Fog'

     8='Other'

     9='Unknown';

6) The number of vehicles variables in the dataset simply records the
total number of vehicles in the crash. 

7) The codes for the “hours driving” variable are included below:

 1-95, actual number of hours driving.

  96='UNK-LEGAL' 

  97='UNK-NOT LEGAL' 

  98='NOT APP' 

  99='UNKNOWN'

For technical Assistance, contact:

Dan Blower

Director, Center for National Truck and Bus Statistics

University of Michigan Transportation Research Institute

(734) 763-6079. 

