HIGHWAY SAFETY INFORMATION SYSTEM GUIDEBOOK FOR THE UTAH STATE DATA FILES

Prepared by
Forrest M. Council
Carolyn Williams
University of North Carolina
Highway Safety Research Center
730 Airport Road, Suite 300
Chapel Hill, NC 27599-3430

Yusuf M. Mohamedshah
LENDIS Corporation
6300 Georgetown Pike
McLean, VA 22101

Prepared for
Federal Highway Administration
Office of Safety and Traffic Operations Research and Development
U.S. Department of Transportation
Washington, DC 20590

August, 2000

TABLE OF CONTENTS

• INTRODUCTION
• DETAILS OF MAJOR FILES
  • The Accident Subfiles
  • VIN Data
  • The Roadlog File
  • Traffic Data
  • The Horizontal Curvature File
  • The Vertical Grade File
  • Issues Related to Merging Files
• UTAH CONTACTS
• COMPOSITE LIST OF VARIABLES
• ACCIDENT FILES
• ACCIDENT SUBFILE
• VEHICLE SUBFILE
• OCCUPANT SUBFILE
• ROADLOG FILE
• CURVE FILE
• GRADE FILE

INTRODUCTION

(NOTE: Changes from the previous edition of the Guidebook are shown in bold and italics.)

Utah's basic highway information system is one of the most complete data base management systems found in any of the HSIS States in terms of the number of files included in the system and the flexibility of output. Since the mid-1980's, the system has either been known as the AHighway Information System@ or the ASR/Route and Traffic System.@ It has been a relational system built on a data base management system named ADABAS. The flexibility of ADABAS allows Utah to "spin off" any number of variables into different file formats. In 2000, the older IBM-based system is being converted to a Client/Server system, run on a UNIX platform, with Informix as the relational database.

The raw files requested for the HSIS include the following:

Accident data:

• Accident subfile
• Vehicle subfile
• Occupant subfile
• Roadlog File (Inventory information)
• Horizontal Curve File
• Vertical Grade File

The decision to request subfiles rather than one large accident file which would capture all accident, vehicle, and occupant data together was made in order to be consistent with other HSIS State formats. For additional consistency, variables related to drivers which did not originally appear on the Vehicle subfile have been copied from the Occupant subfile into that subfile. For ease of use, the three accident subfiles, the Roadlog File, the Horizontal Curve File, and the Grade File have been converted to SAS files.

Raw file data are provided to the Highway Safety Research Center where they are retained as backup information. The documentation (variable listings, definitions, etc.) for these raw files and for the SAS files that are developed from them are available at FHWA offices.

Beginning in 1994, the HSIS system was converted to a relational database for internal use. This database, using a SYBASE system, stores the data received from Utah and other states, and the data files for a given state are linked and manipulated using SQL language. However, this conversion from the original SAS-based system to the newer relational system is somewhat transparent to the end-user of the data since the output files produced by SYBASE for modeling and analysis will be SAS formatted. As in the past, we have continued to produce SAS format libraries for each of the variables in each of the files. Because it is envisioned that the majority of analyses will utilize these SAS files and formats, this Guidebook will concern these SAS files -- their formats, completeness, and quality. While single-variable tables for key variables from each file were published in the past as “Volume II”, this is no longer the case. Users may request specific single- or multiple-variable tables through the HSIS “Data Request” procedure found on this web site.

As noted above, the accident data are in three separate files. The Accident Subfile (including crashes from 1985-97), containing basic information on accident type, location, environment, etc. can be linked with the Vehicle Subfile (which contains information on each vehicle in the crash and each driver) and the Occupant Subfile (which contains information on each occupant in each vehicle and non-occupants such as pedestrians, pedalcyclists, etc.) through use of the accident case number, vehicle number, and year. The Accident and other major files can be linked through use of a route-milepoint system and year with the milepoints for a given route beginning at the State line.

The Roadlog File contains yearly characteristics of 14,000 miles of the road system including shoulder and median information, pavement type and width, lane information, etc. Thus, a separate file exists for 1985, 1986, 1988, 1989, and 1991-1994, and 1996-1997 allowing correct roadway information to be linked with accidents occurring during a given year. (Note that there is no 1995 file due to data linkage problems that will be described later in this narrative.) This file is composed of homogeneous segments of all state and federal-aid roadways which are classified as collectors and above, with a new segment defined each time any of the recorded variables changes. In addition, unlike some other State data bases which use separate files, this Roadlog file also contains traffic information (AADT).

The Horizontal Curve and Vertical Grade Files were generated to produce information on specifics of curvature. This information contains degree of curve, direction of curve, and percent and direction of grade. While a new section in the Roadlog File is begun each time any variable changes (e.g., pavement type), new sections in these files begin only at the beginning and end points of curves and grades. These data can be merged with the Roadlog and Accident Files through the same route-milepoint key. (As described in more detail in the later sections, these files are no longer updated by Utah after 1994.)

Details of the three Accident subfiles, the Roadlog File, the Horizontal Curve File, and the Grade File are presented in the following section.

DETAILS OF MAJOR FILES

The Accident Subfiles

As with the other states in the HSIS system, the Utah accidents that are captured in the system are only those that can be linked with the Roadlog File (the roadway inventory data). In Utah, approximately 85-90 percent of the total reportable crashes in the state are linkable to the inventory information. Non-linkable crashes primarily consist of crashes on roads not mileposted or inventoried where the crash was given a Azone@ location, and crashes which can be located to routes, but not to specific mileposted locations on the routes. Most of these locations (and crashes) are on lower-order county and municipal roads.

The accident data in the State of Utah are coded and edited by staff at Utah DOT. This group locates each of the accidents based on the inputs provided on the report by the investigating officer. Currently, there are thirteen years of accident data in the Utah HSIS files -- 1985-1997. This data set includes between 35,000 and 50,000 accidents per year between 1985 and 1997 , 63,000 -95,000 vehicles per year, and 91,000 - 140,000 occupants/pedestrians per year. Unlike certain other HSIS States, the Occupant subfile contains information on each occupant in the crash, whether injured or not.

To insure data quality, coders within the DOT Division of Safety input the data received from police officers in both urban and rural locations. The same report form is used statewide with a reporting level of $750 total damage and/or personal injury. The coders in Utah DOT headquarters correct the information based on what they see on the form, and actually call the police officer back if they have some question on the accident location or other data. Thus, the feeling is that location information and other variables are approximately 95 percent correct in the rural areas. Re-verification checks for samples of these variables are conducted to check the accuracy. Information provided by the local police is not quite as accurate as that provided by the state police, but it is still considered to be quite accurate.

Based on examination of single- versus multi-vehicle crash percentages and on the fact that city police agencies investigate over 50 percent of the crashes, it appears that the Utah files are slightly weighted toward urban crashes, although not as much as certain other HSIS States.

To investigate data completeness and accuracy, Utah staff was questioned about their experience with files, and a series of single variable tabulations of more than 70 key variables were produced and examined for each year of data in the system. This allows HSIS staff to look for both significant percentages of uncoded or Aunknown@ data within each variable and for changes in individual variable codes across time. When problems are found, Utah staff is contacted to provide an explanation or to conduct further internal checks. The variables examined ranged from various accident descriptors including accident type, day of week, number of vehicles involved, accident severity, road surface, object struck, and investigating agency; to vehicle-related variables involving travel and impact speed, contributing factors and vehicle type; to occupant variables related to safety equipment use, age, sex, and alcohol use.

The quality control checks indicated some problems with a number of variables. However, the problems identified were generally minor in nature, concerned variables which will not be used very often in FHWA analyses, or were problems which can be overcome to some extent by knowledge of how the data behave.

The problems identified can be categorized into a number of categories including (1) variables which have a relatively high percentage of uncoded data; (2) changes in the 1986 accident report form and thus the coding of accident variables; (3) variables which appear to have inconsistent coding across years (often the result of the officers switching from leaving the variable uncoded to a Ano@ or Anone@ code); (4) variables which have problems that could well be related to the nature of the codes available to the investigating officer (e.g., the lack of a Anot injured@ code in the AInjured Body Area@ variable); (5) variables which appear to have problems due to possible errors in computation programs within the Utah system (which have subsequently been corrected); and (6) variables that simply represent poor data. In each of these cases where a problem is found, a ANOTE@ is included under the variable description in the SAS format section which follows. As might be expected, in some cases the data in certain variables improve over time. These instances are noted also.

With respect to the second issue, discussions with Utah staff concerning inconsistencies in the data across years brought to light the fact that there were major changes in the accident report form and thus the way certain variables were coded in 1986. For example, within the "Accident Type" variable (and the related Asubsequent event@ variables), there were no instances of "ran-off-road left or right" in 1985, but a suspiciously large number of instances of "ran-off through median" codes for that year. The explanation was that the code now used for "ran-off through median" (09) represented all ran-off-road crashes prior to 1986. The coding was changed early in 1986 to allow categorizing the ran-off-road accidents into right, left, and through median. In other cases (e.g., ATraffic Control Type"), the variable was expanded in 1986 to include additional variables.

To further investigate the accuracy of the data, comparisons were made of pairs of variables in the file which should have been somewhat similar according to their definitions. In addition, the single-variable tabulations were compared to the other HSIS States to see if large differences existed. The comparisons within the files usually indicated a high degree of consistency between similar variables.

Finally, as noted earlier, the majority of driver-related variables appear both on the Vehicle and Occupant subfiles. However, due to the layout of the original raw files, a limited number of variables (e.g., "Driver License Type," "Years Driving Experience,@ "Drivers License Restriction") appear only on the Vehicle subfile. It is noted that since there are some vehicles which would not have drivers or driver data (e.g., parked vehicles or hit-and-run vehicles), the totals at the bottom of tables for the driver variables copied from the Occupant subfile will be slightly lower than the totals for the driver variables in the original Vehicle subfile. For the latter Vehicle-subfile variables, there will be slightly more "not coded" or "unknown" driver codes.

In summary, it appears that there are a number of variables in the Utah accident files in which there are potential errors that the analyst must be aware of. Fortunately, the majority of these variables will not be primary variables in much of the analysis done by FHWA. However, the problems must be kept in mind in any analyses conducted and, as with other Guidebooks, these errors are noted within the format statements for each variable affected.

VIN Data

Supplemental data on a vehicles characteristics (e.g., engine information, air bag presence, wheelbase, etc.) can be developed by decoding the VIN (Vehicle Identification Number) collected on crash forms by police agencies in some states. For Accident/Vehicle Files from 1987- 94, formats for the decoded VIN files were included in the original Guidebooks, and separate VIN files were developed for each year of data for the states of Utah, Illinois, Michigan, and North Carolina. When a vehicle in the Vehicle File had a legitimate VIN, this decoding was done using the VINDICATOR program distributed by the Highway Loss Data Institute of IIHS (the Insurance Institute of Highway Safety). This detailed information could then be linked back with the vehicle file using the Accident Case Number and the Vehicle Number.

Because of the very limited use of the VIN data by researchers and difficulties we encountered with the decoding process after 1994, the VIN variable listing found in the original Guidebooks for these states has been removed from this version. However, HSIS continues to capture the VIN, and has the capabilities of decoding the VIN for users. For Utah, decoded VIN information is available for 1987-96, and VINs for later years can be decoded upon request. For more information on this data, contact the HSIS staff.

The Roadlog File

Of the approximately 50,000 miles of highway in Utah, approximately 14,000 are included in the HSIS Roadlog File for roadway inventory and accident linkage purposes. Annual files exist for 1985, 1986, 1988, 1989, 1991- 1994, and 1996 -1997. Inventory information on either the 1989 or 1991 files are felt to be suitable for use with 1990 accident data. There is no 1995 file in HSIS. Utah staff Are-mileposted@ crashes for 1995, but did not begin the corresponding Roadlog File re-mileposting until 1996. Thus, the 1995 crashes did not link accurately with the 1995 Roadlog File. Inventory information on the 1996 file is felt to be suitable for linkage with the 1995 crash data. (In some multi-year analyses conducted in the past with HSIS data, multiple years B e.g., three years B of crash data have been linked with one year of inventory data, most often the Amiddle year.@ However, this remileposting means that in such analyses, the researcher should NOT link 1994 and earlier crash data with post-1994 inventory data, or 1995 and later crash data should not be linked with pre-1996 roadway inventory data.)

As noted below, Utah staff indicates highest confidence in the accuracy of the inventory data on the 8,200 miles of roadway on the Federal Aid system and non-Federal Aid "State" or AOther Federal@ system. This includes approximately 950 miles of Interstate roadways, approximately 4,900 miles of other State-system roadways (almost all of which is also on the Federal Aid system), approximately 1,760 miles of non-State, Federal-Aid roadway (primarily Federal-Aid county and municipal roadways) and 640 miles of Aother federal@ roads (primarily US Forest roads). Of the 35,000 - 50,000 accidents that occur each year in Utah, it is estimated that approximately 75 percent occur on roads that are on the Federal Aid System or the non-Federal Aid, State system. The table below provides a categorization of all two-way paved mileage in the system (i.e., unpaved mileage and uncoded mileage is included in the AOther@ category).

Much of the information of the inventory system in Utah is an outgrowth of effort related to collecting data for the Highway Performance Monitoring System (HPMS) B both the AUniverse@ and the ASample@ files. The HPMS Sample-file represents approximately 18 percent of the functionally classifiable roadways in the State. This inventory is updated on a three-year cycle. Using the photologging system, some of the variables that are on the HPMS samples are expanded to cover part or all of the universe of roads within the HSIS Roadlog File. This expansion is being done on a year-to-year priority basis, where the priority is based on the importance of a particular variable for use. And as would be expected, variable coding completeness is related to functional classification of roadways, with the lower functional classes (e.g., minor collectors and local roads) exhibiting less complete coding.

Through examinations of the data across years and by working with the Utah staff, the HSIS staff has attempted to define those variables which do and do not fully cover the entire HSIS system. This is difficult at times, since the expansion is a continuing process and the nature and degree of the expansion differs among variables. In some cases, the expansion may be adding additional miles of coding for a given variable. In other, new variables are added to our files in a certain year (e.g., approximately 30 variables were added in HSIS database in 1991). In addition, there are variables in which uncoded data for some HSIS sections (i.e., variables which are only collected on HPMS-sample sections) have been coded as A0" in the files we receive from Utah. Since A0" is sometimes a legitimate value, this can lead to interpretation problems. Where possible, we have attempted to recode the data based on information concerning the location of HPMS-Sample sections (i.e., replace A0" with blanks). Finally, because the HPMS Sample sections are often longer than the HSIS sections, some variables in the HPMS file would not accurately apply to each HSIS sub-section unless corrected. For example, "Number of Signalized Intersections" for an HPMS sample section which covered three Roadlog sections would not apply accurately to each of the Roadlog sections. Fortunately, for most variables of this type, the Utah staff has corrected the values so that they are accurate for the HSIS subsections.

Since this expansion from HPMS data does lead to possible interpretational and use problems with some variables, HSIS staff has examined each variable in the data across all years of the files and has worked with the Utah staff in an attempt to define possible issues. These issues are then documented in a ANOTE@ in the variable description found in the later SAS format section.

In general, variables which are essentially fully coded for all functional classifications include such key variables as county, section length, rural/urban designation, functional system, AADT, number of through lanes, access control, median type, surface/pavement type, shoulder type, national highway system indicator, and others. Most of the remainder of the variables are available for interstate, primary arterial, major collector, and urban-collector across the years, and have less full coding for minor collectors and local roads. Finally, there is a third set which is coded only for HSIS sections which fall within HPMS Sample sections. Specific information on level of coding is noted in the format section for individual variables.

As in the other HSIS states, each record on the HSIS Roadlog File represents a homogeneous section defined by a change in any of the variables existing on the file. Thus, for example, if pavement width changes for one or more sections, new section "beginnings" may be produced or old sections may be combined from one year to the next.

Approximately 70 percent of the computer records (i.e., homogeneous sections)on the file are on the federal-aid system. These federal-aid records represent approximately 56 percent of the total mileage covered in the file. Thus, the sections in the federal-aid system are somewhat shorter than the sections in the non-federal-aid system. This is logical, in that the more detailed inventory (and thus, more Abreaks@ in homogeneous sections) is found on the federal-aid system. Approximately 85 percent of the mileage in the file is rural in nature. Approximately 85 percent of the 14,000 inventoried miles are two-lane sections, and 12 percent of the mileage have four or more lanes. Approximately half of the roadways on the file are low volume in nature with 51 percent of the roadway mileage having AADT's of 500 or less vehicles per day.

To assess data accuracy and completeness, HSIS staff again relied on past analyses with the data, conversations with Utah staff, and computerized and manual comparisons across years of single-variable tabulations for key variables. As noted above, most of the variables on the 14,000 mile inventory file are coded very completely with very few missing values. However, there are high percentages of unknowns for some variables, particularly those which are only coded for HPMS Sample sections, or some expansion of that coding.

Two new variables, RODWYCLS and MVMT, have been created by HSIS staff in the roadway segment file of each of the HSIS states. The RODWYCLS (Roadway Class) variable is based on the combination of rural/urban, access control, number of lanes and median type variables. This variable classifies each roadway segment into one of ten roadway types described in the later AFormat@ section. This variable is also included as a accident-file variable by matching each crash to its corresponding roadway segment. The MVMT variable (Million Vehicle Miles of Travel) is calculated for each segment in the roadway file by multiplying the segment length, AADT and 365 days in a year, and dividing by one million. Both these variables were created in response to inquires from data users, whose most frequent questions have concerned either crash frequencies or rates (per MVMT) for one or more of these roadway classes.

Finally, added coverage of information occurs in the 1991 file where a number of new variables were added to the Roadlog File. Some refer to engineering details (e.g., thickness of slab), while others appear to be new information on traffic or other variables (e.g., number of signalized intersections). Where available, details are included in the SAS format description.

With respect to data accuracy, as noted above, the inventoried roadways in Utah are composed of three overlapping systems -- the State-controlled system (approximately 5,800 miles, of which 5,600 are also under the Federal-Aid system), additional Federal-Aid highways not on the State-controlled system (approximately 1,700 additional miles, predominantly urban roadways or secondary roadways under township or county control), and approximately 5,700 miles of inventoried roadway which falls neither on the State system nor the Federal Aid system, and which will be referred to as "non-system" roadways. As indicated earlier, the inventory data for these three types of roadways is either based on inventory efforts tied to the requirements of the HPMS system, or on older files in which data were extracted from design drawings or photologs.

Utah staff has the highest confidence in the inventory data which is based on the HPMS effort or the expansion thereof. Thus, all variables on the Federal Aid system (e.g., Interstate, Federal Aid Primary, Federal Aid Secondary and Federal Aid Urban), and all coded variables on the additional State-system routes are felt to be coded very accurately. The inventory data for the remaining 5,700 miles of "non-system" roadway are also felt to be quite good, but not as accurate as the data on the system roads simply because the inventory is not updated as often.

It is noted that as in most state inventory files, the design of the file may lead to some "gaps" in the data in route-oriented analysis. Because the file is section-based, with each section of pavement being represented only once, analyses of roadway types (e.g., all Interstate mileage, all 2-lane mileage) can be conducted accurately, with appropriate accidents being merged to the appropriate sections. However, for analyses involving a specific route, there may be gaps in the data where the route coincides with a second route. Here, the inventory data will only be captured on one of the routes (usually the higher-order route) to prevent double-counting of mileage. For example, since some sections of Interstate I-15 and I-80 coincide, it has been found that the inventory data (and linked accidents) only appear on I-15 in the file. Thus, the analysts doing a route-based analysis on I-80 will have to determine the milepoints at which the routes coincide (gaps in the I-80 file segments which can be linked to a map), and pick up the appropriate data from the I-15 file. Even larger problems would be expected to occur if route-based analysis are attempted for lower-order routes (e.g., state routes which may coincide with US routes). Again, the HSIS files were primarily designed for non-route-based analyses, and will be accurate in the majority of runs conducted.

Traffic Data

As noted earlier, traffic data related to AADT and truck percentages are found on the Roadlog file. These data are based on Utah=s traffic count program. In this program, there are 85 permanent Automatic Traffic Recorders (ATR's) on Interstate and Utah State roads in Utah which are in operation 365 days/year. Of these, 53 ATR's capture volume and vehicle classification counts and 32 ATR's count volume only. These ATR's conform with FHWA's HPMS guidelines. In addition, there are approximately 10 additional ATR=s on roads inside National Parks in Utah that are operated by the National Park Service.

In addition to these permanent counts, Utah collects 48-hour (or more) coverage counts at approximately 1,300 locations per year. Counts on the State-system roadway are done on a 3 to 5 year cycle. Approximately 100 traffic counting machines are utilized to collect traffic data for 7,100 miles of State-system roads in Utah. In terms of coverage, Utah tends to have a better sample coverage of high volume roads compared to lower functional categories. From a purely statistical perspective, a larger sample might be more appropriate for the lower functional classes of roads. However, Utah believes that limited resources for counting should be devoted to the roads that carry a bulk of the traffic. In addition to these coverage counts, approximately 200-300 short-term vehicle classification counts are conducted each year. Finally, to help develop seasonal factors, additional seasonal counts are captured for a week per month or per season.

Short term counts are expanded to AADT estimates using ATR data for roads with similar characteristics, functional class, and volume group. For a year in which no count is made, the previous year=s count for a section is modified by a Agrowth factor@ which is based on changes in counts in the same area or on changes measured by an Aassigned@ (similar) ATR station. In this manner, volume assignments are made to each section of State system roadway each year. Finally, Utah staff also develops estimates of truck percentages and equivalent single axle loadings (ESAL=s) for Aon-system@ roadways based on the classification counts, weigh-in-motion data, interaction with trucking companies, and staff knowledge of trucking routes. Traffic information is entered into the Traffic file as it is being collected, but is transferred to the computerized system, and thus to the Roadlog File, at the end of the year.

With respect to the accuracy of the traffic information, Utah staff indicated that the data are currently being corrected such that errors would probably not be greater than +10 percent for almost all of the State-system sections. (Counts on Anon-system@ county and municipal roads would be less accurate.)

The Horizontal Curvature File

The Horizontal Curvature file has been discontinued by the State of Utah after 1994. According to the State DOT staff, the data are no longer being updated due other priorities in some districts. In addition, the staff has some questions about the accuracy of the original data input into the file. They are currently in the process of verifying data based on as-built plans. To date, even though the staff has questions about the accuracy, the verification activity has not indicated major problems with the data. However, given their concerns, and the fact that the verification process is not complete, we have chosen to leave the pre-1994 data in the HSIS files, and to await the development of the verified file. The following narrative describes the earlier files. However, we feel that the data in the existing (1994 and earlier) files should be used with some caution.

The existing horizontal curvature file contains information on horizontal curves in the State of Utah. It contains 12,976 records in 1994. This file was expanded to the full State-system in 1986 and contains information on curve and tangent sections. (Note that curve-related information does not exist for all Federal-Aid system roads, since some of these roads fall within urban areas not under State DOT control.) The annual files included in the HSIS are for 1987 and 1991-94. It is noted that the sections on this file are divided only at the beginning and end of curves, and thus the number of sections will not match the sections on the Roadlog File which are defined by changes in many other variables.

Approximately 87 percent of the 5,900 miles on this file are tangent sections. This file includes information on the degree of curve, the direction of curve, and the type of roadway on which the tangent and curve sections fall. When compared to Washington and other similar State files, the percentage of curve by degree appears logical. As noted earlier, Utah staff does have some questions concerning the accuracy of these data, and is currently in a data verification process.

The Vertical Grade File

Like the Horizontal Curve File, the Vertical Grade file has been discontinued by the State of Utah after 1994, due to the same updating and data verification issues. (See discussion above.) Again, the data are currently being verified using as-built plans. We have again chosen to leave the 1994 and earlier files in the HSIS system, but again note to the user that the data should be used with caution. The existing file is described in the following narrative.

The vertical curvature file contains information on the percent grade and the direction of grade (either plus or minus). Like the Horizontal Curve File, data are only available for the State-system roads -- approximately 5900 miles of roadway. Currently, the HSIS system contains six files -- 1987, 1989, and 1991-94.

As with the horizontal curvature file, the sections on this file are defined only by breaks in grade. Of the 5900 miles of roadway on this file, approximately 18 percent are characterized by zero percent grade and approximately six percent are characterized by grades steeper than five percent, which suggests that the distribution of grade is quite logical. As noted earlier, Utah staff does have some questions concerning the accuracy of these data, and is currently in a data verification process.

Issues Related to Merging Files

Because the individual Utah files are all extracted from one ADABAS database, all are linkable. As noted above, the accident data are subdivided into three subfiles -- accident, vehicle, and occupant. These subfiles can be linked together using the "case number" variable (i.e., CASENO) present in each of the three files. This variable includes the accident year. When linking the occupant subfile, the additional linking variable "vehicle number" (i.e., VEHNO) must match so that the occupants are associated with the vehicle in which they were traveling. To link the Vehicle subfile with the Accident alone, first sort both subfiles by case number. To link the Occupant file with the other two subfiles, first sort both the Vehicle subfile and Occupant subfile by case number and vehicle number. Next sort the Accident subfile by case number. Alternatively, the separate subfiles can be linked by specifying an SQL JOIN operation with the constraining condition that case number and vehicle number from each table are equal. SQL processing does not require the data to be presorted and the output will not be in any particular sort order unless ORDER BY is specified.

The Accident subfile can then be linked with the Roadlog File using information related to route number and milepost on the route. The actual linkage variables on the Accident file which are used in the merging operation are A_ROUTE (route number) and MILEPOST. The linkage variables on the Roadlog File are BEGMP, ENDMP, and RTE_NBR.

To prepare the Accident subfile for linking with the Roadlog File using a SAS data step process, the analyst must sort both the Accident and the Roadway File into location order by A_ROUTE and MILEPOST on the Accident file and by RTE_NBR and BEGMP on the Roadlog File. Similar sorts would be done with other files to be merged. For the alternative SQL join, the analyst must specify an exact match on A_ROUTE and RTE_NBR from the Accident and Roadlog files and a range match where MILEPOST occurs between BEGMP and ENDMP.

To link the Horizontal Curvature or Grade File with the Accident File, use exactly the same variables as used in linking the Accident and Roadlog Files. Merging the Curve or Grade Files with the Roadlog File is a slightly more complex section-to-section match.

Finally, where appropriate and possible, a format which defines categories within a given variable has been developed for HSIS SAS variables. These categories are shown in the pages below. If you are a SAS user and wish to receive a formatting program which includes these SAS formats (with linkage to the pertinent variable name), please request these from the HSIS staff who provides the data file to you.

UTAH CONTACTS

State computer files and accident information -- Rick Julio (801-965-4268) -- Mr. Julio is our main contact within the State of Utah when questions arise concerning both the data files in general and specific information related to the accident and roadway variables. He is the Roadway Geometric Engineer in the Utah DOT, and has oversight responsibilities for roadway and accident files.

Traffic Engineering information B Mack Christensen (801-965-4264) is the Traffic and Safety Studies Engineer with the Utah DOT. He can provide additional information on uses of the Utah data in safety studies, particularly as related to traffic engineering studies.

COMPOSITE LIST OF VARIABLES FOR ALL UTAH HSIS FILES

LIST OF VARIABLES FOR UTAH ACCIDENT SUBFILE

SAS FORMAT DEFINITIONS FOR VARIABLES FROM THE UTAH ACCIDENT SUBFILE

NOTE: SAS variable names and explanatory names are shown above each listing (See Discussion for information on SAS formats.).

A_ROUTE

ACCIDENT ROUTE NUMBER

NON-LABELED VARIABLE -- Used in file linkage.

A_RTCODE

ROUTE CODE

NON-LABELED VARIABLE -- First character of A_ROUTE. An 'A' designates a State-controlled route.

ACCTYPE

TYPE OF ACCIDENT (FIRST EVENT)

'1' = 'MV--PEDESTRIAN'

Motor vehicle -- Pedestrian

'2' = 'MV--MV'

Motor vehicle -- Motor vehicle

'3' = 'MV--TRAIN'

Motor vehicle -- Train

'4' = 'MV--BICYCLE'

Motor vehicle -- Bicycle

'5' = 'MV--WILD ANIMAL'

Motor vehicle -- Animal (wild)

'6' = 'MV--FIXED OBJ'

Motor vehicle -- Fixed object (mountains, snowbank)

'7' = 'MV--OTR OBJECT'

Motor vehicle -- Other object (boulders, trailers)

'8' = 'OVERTURN IN ROAD'

Overturned in roadway

'9' = 'RAN OFF THRU MED'

Ran off roadway -- thru median

'R' = 'RAN OFF - RIGHT'

Ran off road -- right

'L' = 'RAN OFF - LEFT'

Ran off road -- left

'A','0' = 'OTHER NON-COLL'

Other non-collision

'D' = 'MV--ANIMAL,DOMES'

Motor vehicle -- animal (domestic)

' ' = 'NOT CODED'

 

OTHER = 'ERROR/OTHER CODE'

NOTE: (1) This variable represents the first event in the crash sequence. For subsequent events see also EVENT2 (first subsequent event), EVENT3 (second subsequent event), and EVENT4 (third subsequent event). It is noted that EVENT3 (second subsequent event) contains little information - 90% - 100% of the data are 'NOT CODED'. For similar reasons, EVENT4 was deleted after 1994.

(2) Coding for this variable changed early in 1986, when the "letter" codes were added. Thus, the 1985/early 1986 data for code '9' (RAN OFF THRU MED) are in error. In 1985 and part of 1986, this code represents the total number of ran-off-road crashes.

ACCYR

ACCIDENT YEAR

NON-LABELED VARIABLE -- Year of accident - YYYY

AGENCY

INVESTIGATING OFFICER

'1' = 'UT HWY PATROL'

Utah Highway Patrol (insert badge number)

'2' = 'CITY POLICE'

City police

'3' = 'SHERIFF'

Sheriff

'4' = 'SECURITY FORCE'

University, forest, security forces

' ' = 'NOT CODED'

 

OTHER = 'ERROR/OTHER CODE'

CASENO

ACCIDENT YEAR + CASE NUMBER

• NON-LABELED VARIABLE -- FORMAT 'YYYYNNNNN'
• WHERE YYYY=YEAR
• NNNNN=CASENUM

COL_TYPE

COLLISION TYPE BASED ON VEHICLE MOTIONS

'01' = 'HEAD ON'

Opposite directions, both straight, head on

'02' = 'OPP, TURN L=

Opposite directions, one straight, one turning left

'03' = 'SAME, REAREND'

Same direction, both straight, rear end

'04' = 'TURN R/REAREND'

Same direction, one straight, one turning right, rear end

'05' = 'TURN L/REAREND'

Same direction, one straight, one turning left, rear end

'06' = 'OPP, SIDESWIPE'

Opposite directions, both straight, side-swipe

'07' = 'SAME, SIDESWIPE'

Same direction, both straight, side- swipe

'08' = 'SAME, TURN R'

Same direction, one straight, one right turn

'09' = 'SAME, TURN L'

Same direction, one straight, one left turn

'10' = 'SAME, 2 TURN L'

Same direction, both left turn

'11' = 'APPROACH ANGLE'

Both straight, approaching at angle

'12' = 'R APPR., TURN R'

One straight, one approaching from right, turning right

'13' = 'L APPR., TURN L'

One straight, one approaching from left, turning left

'14' = 'R APPR., TURN L'

One straight, one approaching from right, turning left

'15' = 'OPP, BOTH L'

Opposite direction, both turning left

*'16' = 'SAME, TURN R & L'

Same directions, one turning left, one turning right

'17' = 'ANY SINGLE VEH'

Single vehicle

'18' = 'BACKING'

Backing

'19' = 'SAME,BOTH TURN R'

Same direction, both turning right

'20' = 'ANGLE, 2 TURN R'

Approaching at angle, both turning right

'21' = 'ANGLE, 2 TURN L'

Approaching at angle, both turning left

'22' = 'STRAIGHT & UTURN'

One straight, one U-turn

'23' = 'OPP, TURN R & L'

Opposite directions, one turning left, one turning right

'24' = 'L APPR., TURN R'

One straight, one approaching from left, turning right

'25' = 'ANGLE, L AND R'

Approaching at angle, one turning left, one turning right

'26' = 'HIT PARKED VEH'

One moving , one parked

'27' = 'HIT AND RUN'

Vehicle left accident scene

'30' = 'OTHER'

Other

' ' = 'NOT CODED'

 

OTHER = 'ERROR/OTHER CODE'

*New codes added in 1995.

NOTE: It appears that coding for certain categories in this variable changed in 1993 and later years. Categories 04-07 increased while 08-09 decreased.

DAYMTH

DAY OF MONTH

NON-LABELED VARIABLE - Accident date of the month.

EVENT2

FIRST SUBSEQUENT EVENT

EVENT3

SECOND SUBSEQUENT EVENT

EVENT4

THIRD SUBSEQUENT EVENT

'1' = 'MV--PEDESTRIAN'

Mv -- pedestrian

'2' = 'MV--MV'

Mv -- Mv

'3' = 'MV--TRAIN'

Mv -- train

'4' = 'MV--BICYCLE'

Mv -- bicycle

'5' = 'MV--WILD ANIMAL'

Mv -- Animal (wild)

'6' = 'MV--FIXED OBJ'

Mv -- Fixed object (mountains, snowbank)

'7' = 'MV--OTR OBJECT'

Mv -- Other object (boulders, trailers)

'8' = 'OVERTURN IN ROAD'

Overturned in roadway

'9' = 'RAN OFF THRU MED'

Ran off roadway - thru median

'R' = 'RAN OFF - RIGHT'

Ran off road - right

'L' = 'RAN OFF - LEFT'

Ran off road - left

'A','O' = 'OTHER NON-COLL'

Other non-collision

'D' = 'MV--ANIMAL,DOMES'

Mv -- Animal (domestic)

' ' = 'NOT CODED'

 

OTHER = 'ERROR/OTHER CODE'

NOTE: (1) These variables represent the second, third, and fourth events in the crash sequence. For first event, see ACC_TYP. However, Subsequent EVENT4 is not collected after 1994, and the included data may be erroneous. Do not use as an analysis variable.

(2) Coding for this variable changed in 1986, when the "letter" codes were added. Thus, the 1985 data for code '9' (RAN OFF THRU MED) are in error. In 1985, this code represents the total number of ran-off-road crashes.

HOUR

TIME OF ACCIDENT (HOUR ONLY)

• 00 = '12 MID - 12:59AM'
• 01 = '1 AM - 1:59 AM'
• 02 = '2 AM - 2:59 AM'
• 03 = '3 AM - 3:59 AM'
• 04 = '4 AM - 4:59 AM'
• 05 = '5 AM - 5:59 AM'
• 06 = '6 AM - 6:59 AM'
• 07 = '7 AM - 7:59 AM'
• 08 = '8 AM - 8:59 AM'
• 09 = '9 AM - 9:59 AM'
• 10 = '10 AM - 10:59 AM'
• 11 = '11 AM - 11:59 AM'
• 12 = '12 NOON-12:59 PM'
• 13 = '1 PM - 1:59 PM'
• 14 = '2 PM - 2:59 PM'
• 15 = '3 PM - 3:59 PM'
• 16 = '4 PM - 4:59 PM'
• 17 = '5 PM - 5:59 PM'
• 18 = '6 PM - 6:59 PM'
• 19 = '7 PM - 7:59 PM'
• 20 = '8 PM - 8:59 PM'
• 21 = '9 PM - 9:59 PM'
• 22 = '10 PM - 10:59 PM'
• 23 = '11 PM - 11:59 PM'
• 24 = 'INVALID VALUE'
• . = 'NOT CODED'
• OTHER = 'ERROR/OTHER CODE'

NOTE: The value '24' is an invalid value. According to Utah staff, the small number of cases could possible be combined with data coded as '0'.

INT_TYPE

INTERSECTION TYPE

NOTE: New variable added in 1995.

'A','a' = '4-LEGGED'

Four legged intersection

'B' - 'E', 'b' - 'e' = 'TEE'

Tee intersection

'F' - 'I', 'f' - 'i' = '5-LEGGED'

Five legged intersection

'J' - 'M', 'j' - 'm' = 'WYE'

Wye intersection

' ' = 'NOT CODED'

 

OTHER = 'ERROR/OTHER CODE'

LIGHT

LIGHT CONDITIONS

'1' = 'DAYLIGHT'

Daylight

'2' = 'DAWN'

Dawn

'3' = 'DARK, NOT LITED'

Dark - no street lights

'4' = 'DARK/ ST LIGHTS'

Dark - street lights

'5' = 'DUSK'

Dusk

' ' = 'NOT CODED'

 

OTHER = 'ERROR/OTHER CODE'

NOTE: Prior to 1988, code 2 was "DAWN OR DUSK". Code 5 added in 1989.

LOCALITY

KIND OF LOCALITY

'1' = 'INDUSTRIAL'

Manufacturing/industrial

'2' = 'SHOP/ BUSINESS'

Shopping/business

'3' = 'RESIDENTIAL'

Residential

'4' = 'SCHOOL'

School

'5' = 'FARMS & FIELDS'

Farms and fields

'6' = 'OPEN COUNTRY'

Open country

'7' = 'CHURCH'

Church

'8' = 'PLAYGROUND'

Playground

'9' = 'RAILROAD XING'

Railroad crossing

LOCATN

LOCATION

'1' = 'ST RD & INT'

State Roads and Interstates

'2' = 'CITY PARK'

City Park

'3' = 'RURAL'

Rural

'4' = 'URBAN'

Urban

'5' = 'PRIV PROP/PARKNG'

Private Property/Parking Lot

'6' = 'REST AREA/POE'

Rest Area/Point of Entry

'7' = 'INDIAN LAND'

Indian Land

'8' = 'ST & FED PROP'

State and Federal Property

'9' = 'FOREST LAND'

Forest Land

NOTE: New variable added in 1995.

MILEPOST

ACCIDENT MILEPOST

NON-LABELED VARIABLE

MONTH

MONTH

• 01 = 'JAN'
• 02 = 'FEB'
• 03 = 'MAR'
• 04 = 'APR'
• 05 = 'MAY'
• 06 = 'JUN'
• 07 = 'JUL'
• 08 = 'AUG'
• 09 = 'SEPT'
• 10 = 'OCT'
• 11 = 'NOV'
• 12 = 'DEC'
• . = 'NOT CODED'
• OTHER = 'ERROR/OTHER CODE'

NOCC_FAT

TOTAL NON-OCCUPANTS KILLED

• 0 - 10 = ACTUAL NUMBER
• 11 - 20 = '11 - 20'
• 21 - 50 = '21 - 50'

NOCC_INJ

TOTAL NON-OCCUPANTS INJURED

• 0 - 10 = ACTUAL NUMBER
• 11 - 20 = '11 - 20'
• 21 - 50 = '21 - 50'

NOCCSEV

SUM OF NON-OCCUPANT SEVERITY

NON-LABELED VARIABLE

NOTE: This variable is computed by adding the individual severity codes (e.g., value of 7 if one non-occ killed (5) and another non-occ class c injury (2)). This variable is not available after 1995.

NUMVEHS

NUMBER OF VEHICLES IN CRASH

• 0 - 4 = ACTUAL NUMBER OF VEHICLES IN CRASH
• 5 - 10 = '5 - 10'
• 11 - 20 = '11 - 20'
• 21 - 50 = '21 - 50'
• . = 'NOT CODED'
• OTHER = 'ERROR/OTHER CODE'

NOTE: This variable contains certain cases coded as "0" vehicles. These are errors.

OBJECT1

OBJECT STRUCK

1986 and Later Codes

'A' = 'GUARDRAIL'

Guardrail

'B' = 'GUARDRAIL END'

Guardrail/end sector

'C' = 'UTILITY POLE'

Utility pole

'D' = 'SIGN POST'

Sign post

'E' = 'DELINEATOR POST'

Delineation post

'F' = 'BRIDGE/CULVERT'

Bridge culvert or other highway structure

'G' = 'CURB'

Curb

'H' = 'SAFETY ISLAND'

Curb or safety island

'I' = 'FENCE'

Fence

'J' = 'CONCRETE BAR'

Rigid barrier (concrete)

'K' = 'CRASH CUSHION'

Crash attenuator

'L' = 'EMBANKMENT'

Dirt embankment/ditch/mountain side

'M' = 'WILD ANIMAL'

Wild animal

'N' = 'DOMES. ANIMAL'

Domestic animal

'O' = 'SNOW BANK'

Snow embankment

'P' = 'MAILBOX'

Mailbox or fire hydrant

'Q' = 'CHANNELIZER'

Traffic channelization device

'R' = 'TREE SCRUB'

Tree/Shrubbery

'S' = 'BUILDING'

Building/other structure (wall)

'T' = 'OTHER OBJECT'

Other object

'U' = 'BAR END TRMT'

End treatment for concrete barrier

Pre-1986 Codes

'0' = '*N.S. (PRE-1986)'

 

'1' = '*GUARDRAIL'

Guardrail

'2' = '*UTILITY POLE'

Utility pole

'3' = '*SIGN POST'

Sign post

'4' = '*DELINEATOR POST'

Delineator post

'5' = '*BRIDGE/CULVERT'

Bridge culvert or other highway structure

'6' = '*CURB'

Curb

'7' = '*SAFETY ISLAND'

Curb or safety island

'8' = '*FENCE'

Fence

'9' = '*OTR (TREE,BLDG)'

Other (tree, building)

' ' = 'NOT CODED'

 

OTHER = 'ERROR/OTHER CODE'

NOTE: (1) The coding for this variable changed in 1986. As shown above, the pre-1986 codes were numeric and values will be preceded by an asterisk when they appear in tables. In addition, code 'U' (End treatment for concrete barrier) was added in 1992.

(2) This variable is an Accident File variable through 1994, and then is switched to a Vehicle File variable starting in 1995.

OCCSEV

SUM OF OCCUPANT SEVERITY CODES

NON-LABELED VARIABLE

NOTE: This variable is not available for 1995 and later.

PEDACT

PEDESTRIAN/BICYCLIST ACTION

'00' = 'NOT STATED'

Not stated

'01' = 'CROSSING W/SGNAL'

Crossing at intersection - with signal

'02' = 'XING AGAINST/SIG'

Crossing at intersection - against signal

'03' = 'XING/NO SIGNAL'

Crossing at intersection - no signal

'04' = 'XING/DIAGONALLY'

Crossing at intersection - diagonally

'05' = 'XING/NON-NTERSCT'

Crossing not at intersection

'06' = 'WALK W/TRAFFIC'

Walking in roadway - with traffic

'07' = 'WALK AGAINST/TRF'

Walking in roadway - against traffic

'08' = 'STAND ON MEDIAN'

Standing on median island in crosswalk

'09' = 'OTR STAND IN RD'

Other standing in roadway

'10' = 'GET ON/OFF BUS'

Getting on/off bus

'11' = 'GET ON/OFF OTHER'

Getting on/off other vehicle

'12' = 'PUSH/WORK ON VEH'

Pushing/working on vehicle in roadway

'13' = 'OTR WORK IN RDWY'

Other working in roadway

'14' = 'PLAY IN ROADWAY'

Playing on roadway

'15' = 'BEHIND PARKED'

Coming from behind parked cars

'16' = 'HITCHING ON VEH'

Hitching on vehicle

'17' = 'LYING IN ROADWAY'

Lying in roadway

'18' = 'VENDING IN RDWY'

Vending in roadway

'19' = 'OTHER IN ROADWAY'

Other in roadway

'20' = 'NO IN ROADWAY'

Not in roadway

*'21' = 'RDING WITH TRAFF'

Riding in roadway with traffic

*'22' = 'RDING AGNST TRAF'

Riding in roadway against traffic

*'23' = 'WATNG T/FRM SCHL'

Waiting on route to or from school

*'24' = 'WALKNG ON SDWLK'

Walking on sidewalk

*'25' = 'RDING ON SDWLK'

Riding on sidewalk

' ' = 'NOT CODED'

 

OTHER = 'ERROR/OTHER CODE'

*New codes added in 1994.

NOTE: This variable is in the Accident File through 1994, and is converted to an Occupant File variable for 1995 and later.

PROPDAM

TOTAL PROPERTY DAMAGE

NON-LABELED VARIABLE - the actual or estimated property damage cost for the entire accident.

NOTE: Data are no longer coded since 1989.

RAMP_NO

RAMP NUMBER

NON-LABELED VARIABLE

NOTE: Ramp number is coded by Utah for interchanges on Interstate highways from a series of maps of interchange configurations. Map documentation can be obtained if needed, but for analysis purposes, any character in this field will signify that the crash occurred on a ramp. However, Utah officials noted that there could be some inconsistencies in the data.

RD_CHAR1

ROAD CHARACTER

'1' = 'STRAIGHT/LEVEL'

Straight road -- level

'2' = 'STRAIGHT/GRADE'

Straight road -- on grade

'3' = 'STRAIGHT/CREST'

Straight road -- hillcrest

'4' = 'CURVE/ LEVEL'

Curve -- level

'5' = 'CURVE ON GRADE'

Curve -- on grade

'6' = 'CURVE ON CREST'

Curve -- hillcrest

'7' = 'STRAIGHT/DIP'

Straight -- dip

'8' = 'CURVE/ DIP'

Curve -- dip

RD_DEF

ROAD DEFECT/CONDITION

'1' = 'HLES,RTS-SHLDER'

Holes or ruts in shoulder

'2' = 'HLES,RTS-RDWY'

Holes, ruts, and bumps in roadway

'3' = 'LOOSE MATERIAL'

Loose material

'4' = 'DARK/OBSTRUCTION'

Obstruction not lighted (darkness)

'5' = 'UNMARK OBSTRUCT.'

Obstruction not marked (daylight)

'6' = 'UNDER CONSTRUCT'

Road under construction

'7' = 'ROADWAY REPAIRS'

Roadway repairs

'8' = 'PREV. ACCIDENT'

Obstruction - Previous accident

'9' = 'OTHER'

Other - Specify in remarks

OTHER = 'ERROR/OTHER CODE'

NOTE: All defects increase in 1987. See Discussion.

RDSURF

ROAD SURFACE CONDITIONS

'1' = 'DRY'

Dry

'2' = 'WET'

Wet

'3' = 'MUDDY'

Muddy

'4' = 'SNOWY'

Snowy

'5' = 'ICY'

Icy

'6' = 'OILY'

Oily

' ' = 'NOT CODED'

 

OTHER = 'ERROR/OTHER CODE'

RODWYCLS

ROADWAY CLASSIFICATION

'01' = 'URB FRWY >= 4 LN'

Urban freeways, four or more lanes

'02' = 'URB FRWY < 4 LN'

Urban freeways, less than 4 lanes

'03' = 'URB 2-LANE ROADS'

Urban two-lane roads

'04' = 'URB ML DV N-FREE'

Urban multi-lane divided, non-freeway

'05' = 'URB ML UND N-FRE'

Urban multilane undivided, non-freeway

'06' = 'RUR FRWY >= 4 LN'

Rural freeways, four or more lanes

'07' = 'RUR FRWY < 4 LN'

Rural freeways, less than 4 lanes

'08' = 'RUR 2-LANE ROADS'

Rural two-lane roads

'09' = 'RUR ML DV N-FREE'

Rural multilane divided, non-freeway

'10' = 'RUR ML UND N-FRE'

Rural Multilane undivided, non-freeway

'99' = 'OTHERS'

Others

NOTE: Created variable added to HSIS accident and roadway inventory files in all states in 1999. See Discussion.

RTTYP_AC

ROUTE TYPE - ACCIDENT

'1' = 'STATE ROADS'

State roads

'3' = 'RURAL NON-STATE'

Rural -- non-state

'4' = 'URBAN NON-STATE'

Urban -- non-state

'5' = 'PRIVATE PROP'

Private property

' ' = 'NOT CODED'

 

OTHER = 'ERROR/OTHER CODE'

NOTE: (1) When an accident occurs on private property (code "5"), no vehicle or occupant information is computerized for cost reasons.

(2) 1985 data for codes '2' and '3' are not consistent with similar data from later years.

(3) Data are no longer coded for 1995 and later years.

SEVERITY

SEVERITY

'1' = 'PROP DAMAGE ONLY'

Property damage only

'2' = 'C POSSIBLE INJ'

Possible injury

'3' = 'B NON-INCAP INJ'

Bruises and abrasions

'4' = 'A INCAPACITATING'

Broken bones or bleeding wounds

'5' = 'K FATAL'

Fatal

TAKEBY

INJURED TAKEN BY

'1' = 'PRIV AMBULANCE'

Ambulance, private

'2' = 'AMBULANCE/FIRE'

Ambulance, fire

'3' = 'PARAMEDICS'

Paramedics

'4' = 'PRIVATE VEHICLE'

Private vehicle

'5' = 'HELICOPTER'

Helicopter

'6' = 'OTHER'

Other

' ' = 'NOT CODED'

 

OTHER = 'ERROR/OTHER CODE'

TIME

TIME OF ACCIDENT

NON-LABELED VARIABLE -- FORMAT HHMM WHERE

• HH=HOURS
• MM=MINUTES

TIMEARR

TIME OF ARRIVAL ON SCENE

NON-LABELED VARIABLE -- FORMAT HHMM WHERE

• HH=HOURS
• MM=MINUTES

TIMECALL

TIME OF CALL

NON-LABELED VARIABLE -- HH:MM (i.e.,'00:00' to '23:59')

NOTE: New variable added in 1995.

TIMENOTE

TIME WHEN DISPATCH WAS NOTIFIED

NON-LABELED VARIABLE -- FORMAT HHMM WHERE

• HH=HOURS
• MM=MINUTES

TOT_INJ

TOTAL OCCUPANTS INJURED

• 0 - 10 = ACTUAL NUMBER
• 11 - 20 = '11 - 20'
• 21 - 50 = '21 - 50'

TOT_KILL

TOTAL OCCUPANTS KILLED

• 0 - 10 = ACTUAL NUMBER
• 11 - 20 = '11 - 20'
• 21 - 50 = '21 - 50'

TOT_OCC

TOTAL NUMBER OF PEOPLE IN ACCIDENT

• 0 - 10 = ACTUAL NUMBER
• 11 - 20 = '11 - 20'
• 21 - 90 = '21 - 90'

NOTE: More than 10 percent of "0" codes in 1987 data. This could be due to a Utah program error.

TREATBY

FIRST AID GIVEN BY

'00','0' = 'UNKNOWN'

Unknown

'01','1' = 'POLICEMAN'

Policeman

'02','2' = 'FIREMAN'

Fireman

'03','3' = 'AMBULANCE PERSON'

Ambulance personnel

'04','4' = 'PARAMEDICS'

Paramedics

'05','5' = 'DOCTOR'

Doctor

'06','6' = 'PRIV INDIVIDUAL'

Private individual

'07','7' = 'HOSPITAL'

Hospital

'08','8' = 'HELICOPTER PEOP'

Helicopter personnel

'09','9' = 'NONE ADMINISTER'

None administer

' ' = 'NOT CODED'

 

OTHER = 'ERROR/OTHER CODE'

NOTE: The 1985 data have leading zeros.

TRF_CNTL

TRAFFIC CONTROL TYPE

'1' = 'OFFICER'

Officer or watchman

'2' = 'FLAGMAN'

Flagman

'3' = 'TRAFFIC SIGNAL'

Traffic signal

'4' = 'FLASHING SIGNAL'

Traffic signal (flashing)

'5' = 'STOP SIGN'

Stop sign

'6' = 'YIELD SIGN'

Yield sign

'7' = 'RAILRD GATES'

Railroad gates or signal

'8' = 'OTHER'

Other (specify)

'9' = 'NO CONTROL'

No control present

'A' = 'WARNING SIGN'

Slow or warning sign

'B' = 'LANE MARKING'

Traffic lanes marked

'C' = 'NO PASSING LANES'

No passing lanes

'D' = 'ONE-WAY STREET'

One-way road or street

'E' = 'CONSTRUCT AREA'

Construction or work area

' ' = 'NOT CODED'

 

OTHER = 'ERROR/OTHER CODE'

NOTE: Letter codes added in 1986; thus 1985 data are not consistent with later years.

VER_CODE

VERIFICATION CODE

NON-LABELED VARIABLE

NOTE: This internal coding variable is used to depict whether or not the coded data have undergone all verification checks by Utah staff. A blank signifies that the data have been verified.

WEATHER

WEATHER

'1' = 'CLEAR'

Clear (or Clear/Cloudy pre-1989)

'2' = 'RAINING'

Raining

'3' = 'SNOWING'

Snowing

'4' = 'FOG'

Fog

'5' = 'DUST'

Dust

'6' = 'MIST'

Mist

'7' = 'SLEETING'

Sleeting

'8' = 'CLOUDY'

Cloudy

'9' = 'WINDSTORM'

Windstorm

' ' = 'NOT CODED'

 

OTHER = 'ERROR/OTHER CODE'

NOTE: Codes 8 and 9 new in 1992.

WEEKDAY

DAY OF WEEK

1 = 'MON'

Monday

2 = 'TUES'

Tuesday

3 = 'WED'

Wednesday

4 = 'THUR'

Thursday

5 = 'FRI'

Friday

6 = 'SAT'

Saturday

7 = 'SUN'

Sunday

0 = 'N.S.'

Not specified

XSECT

INTERSECTION CODE

'A00' = 'A/4-WAY NTERSECT'

4-way intersection

'B00' = 'B/T TO THE WEST'

T to the west

'C00' = 'C/T TO THE EAST'

T to the east

'D00' = 'D/T TO THE SOUTH'

T to the south

'E00' = 'E/T TO THE NORTH'

T to the north

'F00' = 'F/5-WAY,LG TO NW'

5-way, leg to northwest

'G00' = 'G/5-WAY,LG TO NE'

5-way, leg to northeast

'H00' = 'H/5-WAY,LG TO SW'

5-way, leg to southwest

'I00' = 'I/5-WAY,LG TO SE'

5-way, leg to southeast

'J00' = 'J/Y TO THE SOUTH'

Y to the south

'K00' = 'K/Y TO THE NORTH'

Y to the north

'L00' = 'L/Y TO THE WEST'

Y to the west

'M00' = 'M/Y TO THE EAST'

Y to the east

'Z00' = 'Z/OTHER INTER'

Other intersection

NOTE: The above coding is based on the intersection sketch provided by the officer. The orientations are relative to North. Thus, B00 is a T-intersection with the base of the "T" toward the west (left). The data are not available after 1995.

LIST OF VARIABLES FOR UTAH VEHICLE SUBFILE

SAS FORMAT DEFINITIONS FOR VARIABLES FROM THE UTAH VEHICLE SUBFILE

NOTE: (1) SAS variable names and explanatory names are shown above each listing (See Discussion for information on SAS formats.).

(2) For consistency with other State's files and ease of handling, driver-related variables have been included in this Vehicle Subfile as well as in the Occupant Subfile.

ALT_VEH

ALTERED VEHICLE

'1' = 'SUSPENSION'

Suspension altered

'2' = 'BODY'

Body altered

'3' = 'TINTED WINDOWS'

Tinted windows

'4' = 'OTHER'

Other alteration

'5' = 'NONE'

 

'U' = 'UNKNOWN'

 

' ' = 'NOT CODED'

 

OTHER = 'ERROR/OTHER CODE'

NOTE: New variable added in 1995.

CARGO

DESCRIPTION OF CARGO

'A' = 'GENERAL FREIGHT'

General freight

'B' = 'HOUSEHOLD GOODS'

Household goods

'C' = 'HEAVY MACHINERY'

Heavy machinery

'D' = 'MOTOR VEHICLES'

Motor vehicles

'E' = 'GASES IN BULK'

Gases in bulk

'F' = 'LIVESTOCK'

Livestock

'G' = 'SOLIDS IN BULK'

Solids in bulk

'H' = 'LIQUIDS IN BULK'

Liquids in bulk

'I' = 'EXPL/HAZRDS MTRL'

Explosives/hazardous materials

'J' = 'REFRIG FOODS'

Refrigerated foods

'K' = 'EMPTY'

Empty

'L' = 'OTHER'

Other

'U' = 'UNKNOWN'

Unknown

'X' = 'N/A'

Does not apply

' ' = 'NOT CODED'

 

OTHER = 'ERROR/OTHER CODE'

CARR_ADD

CARRIER ADDRESS

NON-LABELED VARIABLE - Description of carrier address or P.O.BOX.

CARR_CTY

CARRIER CITY

NON-LABELED VARIABLE - City name

CARR_NAM

CARRIER NAME

NON-LABELED VARIABLE - Carrier name (company name or mailbox address)

CARR_STE

CARRIER STATE

NON-LABELED VARIABLE - Two-letter postal abbreviation code, i.e, UT(Utah).

CARR_ZIP

CARRIER ZIP CODE

NON-LABELED VARIABLE - Zip code, up to ten characters.

CASENO

ACCIDENT YEAR + CASE NUMBER

• NON-LABELED VARIABLE - FORMAT 'YYYYNNNNN'
• WHERE YYYY=YEAR
• NNNNN=CASE NUMBER

COMM_TRK

COMMERCIAL TRUCK

NON-LABELED VARIABLE

NOTE: This variable was used in earlier years to designate "interstate" (1) or "intrastate" (2) shippers. However, the data are not collected.

CONTRIB1

FIRST CONTRIBUTING CIRCUMSTANCES

CONTRIB2

SECOND CONTRIBUTING CIRCUMSTANCES

CONTRIB3

THIRD CONTRIBUTING CIRCUMSTANCES

'00' = 'NONE'

Did not contribute

'01' = 'SPEEDING'

Speed too fast

'02' = 'FAIL TO YIELD'

Failed to yield right of way

'03' = 'LEFT OF CENTER'

Drove left of center

'04' = 'IMPROP OVERTAKE'

Improper overtaking

'05' = 'RAN STOP SIGN'

Passed stop sign

'06' = 'RAN TRAF SIGNAL'

Disregard traffic signal

'07' = 'TOO CLOSE'

Followed too closely

'08' = 'IMPROPER TURN'

Made improper turn

'09' = 'ALCOHOL'

Had been drinking

'10' = 'DUI - DRUGS'

Under the influence of drugs

'11' = 'POOR EYESIGHT'

Eyesight defective uncorrected

'12' = 'ASLEEP'

Asleep

'13' = 'FATIGUED'

Fatigued

'14' = 'ILL'

Ill

'15' = 'IMPROPER PARK'

Improper parking

'16' = 'IMPROPER LOOK'

Improper lookout

'17' = 'FAIL TO SIGNAL'

Failed to signal

'18' = 'OTHER IMPROPER'

Other improper driving

'19' = 'DEFECT BRAKES'

Brakes defective

'20' = 'HEADLIGHT OUT'

Headlight insufficient or out

'21' = 'HEADLIGHT GLARE'

Headlights glaring

'22' = 'LIGHT DEFECT'

Other lights or reflectors defective

'23' = 'STEER DEFECT'

Steering mechanism defective

'24' = 'TIRE DEFECT'

Tires defective

'25' = 'WINDSHIELD CLOUD'

Windshield not clear

'26' = 'OTR VEH DEFECT'

Other defective condition of vehicle

'27' = 'HIT AND RUN'

Hit and run

'28' = 'DUI-ALCOHOL'

Driving under influence - alcohol

'29' = 'NON-COLL FIRE'

Non-collision (fire)

'30' = 'COLL + FIRE'

Collision (fire)

'40' = 'STOLEN'

Stolen

'41' = 'PHANTOM VEH'

Non-contact vehicle involved

'42' = 'TRK JACKNIFE'

Truck tractor and trailer jacknifed

'43' = 'TRK RUNAWAY'

Truck Downhill "runaway"

'44' = 'TRK CARGO LOSS'

Truck lost cargo

'45' = 'TRK FIRE'

Truck fire

'46' = 'TRK UNITS SEPRTD'

Truck units separated

'47' = 'WRNG SIDE OF RD'

Wrong side of road

'48' = 'WRNG WY 1-WAY ST'

Wrong way on one way street

'49' = 'IMPROPER BACKING'

Improper backing

'50' = 'IMMERSION'

Immersion

'51' = 'TOWED VEH'

Towed vehicle

'52' = 'VEH ROLLNG IN LN'

Vehicle rolling in traffic lane

' ' = 'NOT CODED'

 

OTHER = 'ERROR/OTHER CODE'

NOTE: The proportions of 'not stated' and 'NONE' vary inconsistently from year to year. Officers may be failing to enter the 'none' code when no factor is present. Codes '42'-'45' are added in 1992. Codes '47'-'52' are added in 1995. CONTRIB3 was not collected after 1989.

DAMAGE2

PART DAMAGED #2

DAMAGE3

PART DAMAGED #3

DAMAGE4

PART DAMAGED #4

DAMAGE5

PART DAMAGED #5

DAMAGE6

PART DAMAGED #6

DAMAGE7

PART DAMAGED #7

DAMAGE8

PART DAMAGED #8

DAMAGE9

PART DAMAGED #9

• '0' = 'NO DAMAGE/TOTLED' No damage or total damage
• '1' = 'LEFT FRONT'
• '2' = 'MID FRONT'
• '3' = 'RIGHT FRONT'
• '4' = 'LEFT SIDE'
• '5' = 'MIDDLE'
• '6' = 'RIGHT SIDE'
• '7' = 'LEFT BACK'
• '8' = 'MIDDLE BACK'
• '9' = 'RIGHT BACK'
• ' ' = 'NOT CODED'
• OTHER = 'ERROR/OTHER CODE'

NOTE: See V_DAMAGE for initial damage. Note that '0' can mean either that the vehicle was undamaged or totaled, so use with caution. DAMAGE7-DAMAGE9 were added in 1995.

DIR_TRVL

DIRECTION OF TRAVEL

• '1','n','N' = 'NORTH'
• '2','s','S' = 'SOUTH'
• '3','e','E' = 'EAST'
• '4','w','W' = 'WEST'
• ' ' = 'NOT CODED'
• OTHER = 'ERROR/OTHER CODE'

DR_BDATE

DRIVER DATE OF BIRTH

NON-LABELED VARIABLE - The actual date of birth taken from the driver license or other identification received in the investigation.

DR_EDUC

DRIVER EDUCATION

'1' = 'PUBLIC'

Public

'2' = 'COMMERCIAL'

Commercial

'3' = 'NONE'

None

'4' = 'UNKNOWN'

Unknown

' ' = 'NOT CODED'

 

OTHER = 'ERROR/OTHER CODE'

NOTE: More than 5% is uncoded. It appears that officers may not be entering the 'unknown' code.

DR_EJECT

DRIVER EJECTION

'1' = 'NOT EJECTED'

Not ejected

'2' = 'PARTIALLY EJECT'

Partially ejected

'3' = 'FULLY EJECT'

Fully ejected

' ' = 'NOT CODED'

 

OTHER = 'ERROR/OTHER CODE'

DR_INJAR

DRIVER INJURED BODY AREA

• '1' = 'HEAD'
• '2' = 'FACE'
• '3' = 'NECK'
• '4' = 'CHEST'
• '5' = 'BACK'
• '6' = 'LEG(S)'
• '7' = 'ARM(S)'
• '8' = 'TORSO'
• '9' = 'UNKNOWN'
• ' ' = 'NOT CODED'
• OTHER = 'ERROR/OTHER CODE'

NOTE: Variable uncoded or 'UNKNOWN' in approximate 80% of cases. However, a large proportion of this is due to the lack of a 'NO INJURY' code within this variable.

DR_INJCS

DRIVER INJURY CAUSE

'1' = 'STEERING WHEEL'

Steering wheel

'2' = 'WINDSHIELD/DASH'

Dashboard/windshield

'3' = 'ROOF'

Roof

'4' = 'OTHER INTERIOR'

Other interior

'5' = 'MCYC HANDBARS'

Motorcycle handbars

'6' = 'MCYC GAS TANK'

Motorcycle gas tank

'7' = 'EXT VEH PART'

Exterior vehicle part

'8' = 'EXTERNAL OBJ'

External object

'9' = 'UNKNOWN'

Unknown

' ' = 'NOT CODED'

 

OTHER = 'ERROR/OTHER CODE'

NOTE: Variable uncoded or 'UNKNOWN' in approximate 80% of cases. However, a large proportion of this is due to the lack of a 'NO INJURY' code within this variable.

DRV_AGEM

DRIVER AGE

• 00 = 'UNKNOWN'
• 01 = '1 YEAR'
• 02-04 = '02-04 YRS'
• 05-10 = '05-10 YRS'
• 11-14 = '11-14 YRS'
• 15 = '15 YRS'
• 16 = '16 YRS'
• 17 = '17 YRS'
• 18 = '18 YRS'
• 19 = '19 YRS'
• 20 = '20 YRS'
• 21-25 = '21-25 YRS'
• 26-30 = '26-30 YRS'
• 31-35 = '31-35 YRS'
• 36-45 = '36-45 YRS'
• 46-55 = '46-55 YRS'
• 56-65 = '56-65 YRS'
• 66-99 = '>= 66 YRS'

DRV_BAC

DRIVER ALCOHOL PERCENT

• 000 = '000'
• 001-007 = '0.01 - 0.07'
• 008-009 = '0.08 - 0.09'
• 010-015 = '0.10 - 0.15'
• 016-020 = '0.16 - 0.20'
• 021-030 = '0.21 - 0.30'
• 031-999 = ' >= 0.31'

NOTE: The total proportion of positive BAC's and the portion of '>= 0.31' BAC's are both suspect.

DRV_INJ

DRIVER INJURY TYPE

'1' = 'NO INJURY'

No injury

'2' = 'POSSIBLE INJURY'

Possible injury

'3' = 'BRUISE/ABRASION'

Bruises and abrasions

'4' = 'BLEEDING/BROKEN'

Broken bones/bleeding wounds

'5' = 'FATAL'

Fatal

' ' = 'NOT CODED'

 

OTHER = 'ERROR/OTHER CODE'

DRV_REST

DRIVER SAFETY EQUIPMENT

'1' = 'LAP BELT USED'

Lap belt used

'2' = 'LAP&SHLD USED'

Lap and shoulder belt used

'3' = 'BELTS NOT USED'

Belts not used

'4' = 'NO BELTS INSTALL'

No belts installed

'5' = 'CHILD RESTRAINT'

Child restraint used

*'6' = 'AIR BAG W BELT'

Air bag deployed with belt

*'7' = 'AIR BAG WO BELT'

Air bag deployed with no belt

'8' = 'HELMET WORN'

Helmet worn

'9' = 'EYE PROTEST USED'

Eye protection used

'0' = 'HELMET+EYE USED'

Helmet worn and eye protection used

'A' = 'SHOULDR BLT ONLY'

Shoulder belt only used

'B' = 'OTHER'

Other

'C' = 'UNKNOWN'

Unknown

'G' = 'AIR BAG DPLO<>94'

Air bag deployed (pre-1994)

' ' = 'NOT CODED'

Not coded

OTHER = 'ERROR/OTHER CODE'

*New codes added in 1994.

NOTE: The proportion of belts used appears to be higher than expected, particularly in 1987. As in other States, this could be the effect of occupants providing incorrect information due to the mandatory belt law.

DRV_SEX

DRIVER SEX

• 'U','0' = 'UNKNOWN'
• 'M','1' = 'MALE'
• 'F','2' = 'FEMALE'
• ' ' = 'NOT CODED'
• OTHER = 'ERROR/OTHER CODE'

NOTE: Conflicting raw file documentation. Above coding is correct according to Utah.

DRV_STAT

DRIVER LICENSE STATE

NON-LABELED VARIABLE - State name (e.g.,UT = Utah)- Use with caution.

IMPT_SPD

IMPACT SPEED

• 00 = 'NOT MOVING'
• 01-05 = '1 - 5 MPH'
• 06-10 = '6 - 10 MPH'
• 11-15 = '11 - 15 MPH'
• 16-20 = '16 - 20 MPH'
• 31-35 = '31 - 35 MPH'
• 21-25 = '21 - 25 MPH'
• 26-30 = '26 - 30 MPH'
• 36-40 = '36 - 40 MPH'
• 41-45 = '41 - 45 MPH'
• 46-50 = '46 - 50 MPH'
• 51-55 = '51 - 55 MPH'
• 56-60 = '56 - 60 MPH'
• 61-65 = '61 - 65 MPH'
• 66-70 = '66 - 70 MPH'
• 71-75 = '71 - 75 MPH'
• 76-80 = '76 - 80 MPH'
• 81-85 = '81 - 85 MPH'
• 86-90 = '86 - 90 MPH'
• 91-95 = '91 - 95 MPH'
• 96-99 = '>= 96 MPH'

NOTE: Approximately 50% of the cases are coded as "00" = "NOT MOVING". This could be a valid code for parked or stopped vehicles. However, it is also the default code for uncoded cases. Use with caution.

INSPECT

SAFETY INSPECTION

• '1' = 'INSPECTED'
• '2' = 'NO INSPECTION'
• ' ' = 'NOT CODED'
• OTHER = 'ERROR/OTHER CODE'

NOTE: Uncoded data has increased to 99% in 1987. Do not use as analysis variable. The variable is deleted in 1995.

LICRESTR

DRIVER LICENSE RESTRICTION

'A','00' = 'NO RESTRICTIONS'

No restrictions (see Note)

'B','01','1 ' = 'CORRECTIVE LENS'

Corrective lenses

'02' = 'LEFT MIRROR'

Left or outside rear view mirror (See '11' also)

'03' = 'DAYLIGHT DRIVING'

Daylight driving

'04' = 'NOT TO EXCEED'

Not to exceed 40 m.p.h.

'05' = 'AUTO TRANS'

Automatic transmission

'06' = 'CUSHIONS'

Cushions or mechanical or prosthetic aid

'07' = 'HAND CONTROLS'

Hand controls

'08' = 'POWER STEERING'

Power steering

'09' = 'POWER BRAKES'

Power brakes

'10' = 'SIGNAL LIGHTS'

Signal lights

'11' = 'OUTSIDE MIRRORS'

Outside rear view mirrors (see '02' also)

'12' = 'LIM HRS/NO AIRBR'

Limited hours or vehicle without air brakes

'13' = 'MISC RESTRICT'

Misc. restrictions (i.e. contact in one eye only, employment, CDL Interstate only, etc.)

'14' = 'MEDICAL'

All yearly or periodical medical renewals

'15' = '90-CC CYCLE'

90-cc motorcycle

'16' = '3-WHEEL CYCLE'

3-wheel cycle or 90-cc Motorcycle (see >15' also)

'17' = 'LESS 10,000 LBS'

10,000 lbs or less

'18' = 'NO HAZARD MAT'

No hazardous materials

' ' = 'NOT CODED'

 

OTHER = 'ERROR/OTHER CODE'

NOTE: The proportions of 'NO RESTRICTIONS' and >NOT CODED= data vary dramatically across years. It appears that both mean "no restrictions." In addition, the codes for certain categories were changed and/or combined sometime between 1990 and 1992. While we have attempted to accurately format the data, be aware that some of the coding in those years is suspect.

LICTYP

DRIVER LICENSE TYPE

• '1' = 'REGULAR'
• '2' = 'CHAUFFEUR'
• '3' = 'MOTORCYCLE'
• '4' = 'RESTRICTED'
• ' ' = 'NOT CODED'
• OTHER = 'ERROR/OTHER CODE'

NOTE: Over 5% of the cases are uncoded in 1986 and 1987.

MISCACT1

DRIVER INTENT

'01' = 'GO STRAIGHT'

Go straight ahead

'02' = 'OVERTAKE'

Overtake

'03' = 'MAKE R TURN'

Make right turn

'04' =