7. Usage of the Crash Data Delivery System

The research team used crash data from the UDOT Crash Data Delivery System, both the 2004 and 2005 versions, for this study. Both versions of the databases were accessed via the Internet, using the URLs in the references section of this report. While accessing the CDDS was straightforward, compiling the information for and creating the tables in this report was often challenging. The methods used to generate the tables are described below. Except where noted, the research team used the 2004 CDDS.

7.1 Ranking Intersections by Total Number of Crashes

The following procedure was used to develop the statewide and regional lists. The 2004 CDDS was used to do these rankings:

1. Select the "Intersections" tool (see Figure 7.1).
2. Enter the starting and ending years of the study period.
3. Select a region or, for Region 4, a district.
4. Set the radius to 500 ft.
5. To limit the search time, enter a minimum number of accidents - the tool will find all intersections having a number of crashes greater than or equal to (not greater than) this number.
6. Hit "Search" and wait for the results.
7. Copy the resultant table and paste it into a spreadsheet, such as Excel (save the spreadsheet).
8. In Excel, sort the table on the number of crashes, producing a ranked list. Refer to this as Table A.
9. Reduce the size of Table A by deleting all crashes occurring on Interstate freeways and on federal-aid roads (i.e., those not occurring on state routes).

Figure 7.1 Intersection Search Example: Region 1 Crashes, 1994-2003

Three types of intersections appear in Table A from step #8:

• Intersections between state routes and other state routes.
• Intersections between state routes and federal-aid routes.
• Intersections between state routes and other roads (non-state; non-federal-aid).

Crashes can occur on all approaches to an intersection. Therefore, the full complement of crash activity at any given intersection takes into account the crashes occurring on both intersecting roads. The CDDS includes information on all motor vehicle crashes occurring on all state and federal-aid routes, but not "other" roads. The next steps in the analysis, therefore, are to tabulate the crashes on the intersecting state and federal-aid routes, then add these values to the crashes that occurred on the state routes. One approach is as follows:

10. In Table A, distinguish among the three intersection types listed above. This may involve adding a column for "intersecting route," then entering a state route number, federal-aid route code, or a blank cell for an "other" road. The names and route numbers of the intersecting roads are found in the "Point Description" column of Table A.
11. For intersections between two state routes, return to the CDDS home page and select the "SR Intersections" tool.
12. Repeat steps #2-6. Using information from Table A, enter a minimum number of accidents that will be conducive to producing a list of intersections with a large number of crashes.
13. Hit "Search" and wait for the results.
14. Copy the resultant table and paste it into a different Excel spreadsheet (save the spreadsheet).
15. In Excel, sort the table on the number of crashes, producing a ranked list. Refer to this as Table B. The analyst now has crash totals for intersections between two state routes.
16. To get crash totals for intersections between state and federal-aid routes, return to the CDDS home page and select the "Advanced Search" tool.
17. Based on the data in Table A, select federal-aid routes at intersections at which the number of crashes is fairly large. Use the data in Table B to restrict the selections.
18. In Advanced Search, select the following search fields: Route_Num and Milepoint.
19. Hit "Build Search," producing a new view with a new set of criteria.
20. Enter the starting and ending years of the study period.
21. Scroll down and select a federal-aid route, based on the results of step #17.
22. Enter the starting and ending milepoints - each value should be the milepoint of an intersection + 0.09 mi (for a 500-ft radius of influence).
23. Hit "Search" and wait for the results. Once the resultant table appears, scroll down to see the total number of crashes. Record the value, or, add it to the number of crashes for the intersecting state route, and enter the sum, along with the intersection information, into Table B.
24. Return to the "Select Criteria" view and select another federal-aid route.
25. Enter the milepoints, as in step #22, and execute step #23.
26. Repeat steps #22 and 23 until all of the federal-aid routes chosen in step #17 have been covered.
27. Table B should now contain crash totals for state route-state route and state route-federal aid route intersections.
28. Finally, copy the crash totals from the state route-other road intersections listed in Table A into Table B.
29. Sort Table B on the crash totals, producing a ranked list.

To develop a statewide list, combine the three regional and three district tables, then do an overall sort. Regarding step #5, the user may want to enter different values until the list reaches its desired length. In Figure 6.3, the inputs shown produce a preliminary list of the 25 intersections having the most crashes in Region 1.

7.2 Ranking Intersections by Crash Severity Scores

There are, potentially, several ways to develop ranked lists of intersections based on crash severity scores. None of the CDDS intersection tools compute severity scores directly, so the analyst must develop an approach. The challenge is to find intersections that have numerous severe crashes. In Regions 1, 2 and 3, and the Cedar City District of Region 4, there were plenty of intersections that had fatal and/or multiple incapacitating injury crashes during the 10- and 3-year study periods. In the Price and Richfield districts, the number of intersections having severe crashes was limited, so it was necessary to also consider bruises-abrasions crashes. The following procedure - not necessarily the only possible approach - was used to develop the severity scores lists. The 2004 CDDS was used:

1. Select the "Advanced Search" tool (see Figure 7.2).
2. Select the following search fields: Region/District and Severity.
3. Hit "Build Search," producing a new view with a new set of criteria.
4. Enter the starting and ending years of the study period.
5. Select a region or district.
6. Select a severity - for the first iteration, select "Fatal."
7. Hit "Search" and wait for the results.
8. Copy the resultant table and paste it into a spreadsheet, such as Excel.
9. Remove all crashes occurring along Interstate and Federal-aid routes (i.e., crashes not occurring on state routes).
10. In Excel, add a column to the spreadsheet and insert a "5" into each row - these values indicate that the crashes are fatal. Alternatively, each row can be highlighted with a distinguishing color.
11. Save the spreadsheet - refer to this as Table C - and return to "Advanced Search."
12. Reselect the search fields Region/District and Severity.
13. Hit "Build Search."
14. Enter the starting and ending years.
15. Select the same region or district.
16. Select a severity - for the second iteration, select "Broken Bones or Bleeding Wounds."
17. Hit "Search" and wait for the results.
18. Copy the resultant table and paste it into Table C.
19. As in #9, distinguish each of these crashes with either a "4" or a different color.
20. Sort Table C by route number and milepoint.

Figure 7.2. Advanced Search Example: Crash Severity by Region or District

As mentioned above, for the Price and Richfield districts, a third iteration on "Bruises and Abrasions" may be necessary. The analyst now has a spreadsheet that includes the fatal, incapacitating injury, and possibly bruises and abrasions crashes that occurred in the given region or district during the study period. The next step is to identify those crashes that occurred within the functional area of an intersection. To reduce the number of locations being considered, it may be useful to examine the routes and milepoints for clusters of crashes. The analyst may define a "cluster" as one or more fatal, or five or more incapacitating injury crashes, for example. For a 500-ft radius of influence, crashes occurring within 0.18 mi of each other may be associated with the same intersection. Single or few crashes at isolated locations may not need to be considered. The analyst should maintain an approximate count of the number of clusters. Once the number exceeds some minimum value, such as 50 or 100, then isolated crashes can be removed from the spreadsheet. Then, to identify intersections:

21. Modify Table C to highlight clusters of fatal and broken bones-bleeding wounds crashes, as discussed above.
22. Take special note of the routes along which the crash clusters occurred.
23. Return to the "Intersections" tool.
24. Enter the starting and ending years.
25. Select a route based on information from the modified spreadsheet.
26. Leave the starting (0) and ending (1000) milepoints as is to ensure that the entire route is searched.
27. Set the radius to 500 ft.
28. Set the "Total accidents greater than" criterion equal to zero to ensure that all crashes along the given route are included.
29. Hit "Search" and wait for the results.
30. The resultant table provides a list of all of the intersections along the state route. Compare the milepoints in this table with those in the modified spreadsheet. Eliminate crash clusters that are not within 0.09 mi of an intersection. It would be useful to enter the milepoints and name(s) of the selected cross-streets into Table C.

The purpose of the final set of steps is to determine the severity scores for the selected intersections. Following step #29, the analyst should have a spreadsheet with fatal and incapacitating injury crashes clustered around intersections along state routes.

31. Return to the "Advanced Search" tool (see Figure 7.3).
32. Select the following search fields: Route_Num (route number), Milepoint, and Severity.
33. Hit "Build Search," and produce a new view with a new set of criteria.
34. Enter the starting and ending years of the study period.
35. Select a route.
36. Enter the starting and ending milepoints - each value should be the milepoint of an intersection + 0.09 mi (for a 500-ft radius of influence).
37. Select a severity - start with "No Injury."
38. Hit "Search" and wait for the results.
39. Once the resultant table appears, scroll down to see the total number of crashes of the selected severity for the given intersection. Record the value for future application.
40. Return to the "Select Criteria" view and select another severity - for the second iteration, choose "Possible Injury."
41. Hit "Search" and wait for the results.
42. Repeat step #38, record the value, then repeat steps #36-40 for the other three crash severities.

As in section 7.1, it is necessary to find the crash severity scores on the intersecting cross-streets. As before, identify the intersecting state route or federal-aid route and milepoint - this information should be available from step #30. For intersections with "other roads," there is no further work to do, and the results obtained in step #41 are final. In "Advanced Search," repeat steps #34-40 for the intersecting state routes and federal-aid routes.

Once the preceding procedure has been completed, the analyst should have the total number of crashes, for each of five severities, for the study intersection, and its cross-street. The analyst can now compute the severity score for the intersection. Steps #31-42, along with the procedure in the paragraph following step #42, must be repeated for every intersection in the spreadsheet. Once this work is done, the analyst can perform a sort on the severity scores.

Figure 7.3. Advanced Search Select Criteria Example: SR 171 at milepoint 10.54

7.3 Locating Intersections with Fatal Crashes

To find intersections having one or more fatal crashes during a given study period, the research team used the "Advanced Search" tool, as in steps #1-9 from section 7.2. Then, to eliminate fatal crashes not occurring at intersections, apply steps #23-30 from section 7.2. The analyst should now have a spreadsheet (Table D) that contains intersections at which fatal crashes occurred. By grouping the routes and milepoints, the analyst should begin to observe intersections at which multiple fatal crashes occurred. To complete the tabulation, it is necessary to identify fatal crashes on the intersecting cross-streets. The procedure described in the paragraph following step #42 in section 7.2 can be applied, except that the analyst is searching for fatal crashes only. Note that all of this information may have been compiled during the procedures described in section 7.2. There may be no need to repeat these steps, but the analyst may be interested in intersections in addition to those summarized in section 7.2.

7.4 Crash Rates

Crash rates for state route-state route intersections can be tabulated using the 2005 version of the CDDS:

1. On the home page, select "Accidents" within the "Applications" menu.
2. Click "Intersection" in the upper menu.
3. Enter the starting and ending years of the study period.
4. Select a region or, for Region 4, a district.
5. Set the radius to 500 ft.
6. Set the "Total accidents greater than" to 2 - this will ensure that crash rates at intersections experiencing no more than one crash during the study period are excluded.
7. Leave the other inputs as is, click "Search," and wait for the results.
8. Column 8 of the resultant table features the crash rate at each state route-state route intersection in the study region.
9. Copy the table and paste into a spreadsheet. Refer to this as Table E.
10. In the spreadsheet, sort the tabulation on the crash rate. The sort produces a ranked list.

The analyst should be aware that the CDDS crash rate is the total number of crashes at the study intersection during the study period divided by the total entering traffic volume during the most recent year, multiplied by one million. The CDDS crash rate, therefore, is accurate only if the study period is one year in length. To properly adjust the rate, one approach is to divide the CDDS crash rate by the length of the study period in years, as follows:

7.5 Non-Signalized Intersections

The research team used the 2005 version of the CDDS to compile crash statistics at non-signalized intersections. The 2005 CDDS featured a "Point of Interests" tool, not available in the 2004 CDDS that made the identification of the traffic control type at each intersection straightforward. To rank non-signalized intersections according to the number of crashes and crash severity scores, the procedures described in Sections 7.1 and 7.2 should be applied. An additional component of the analysis, though, is the traffic control type at each intersection. This can be determined as follows:

1. On the home page, select "Points of Interest Admin" within the Applications menu.
2. Select a route number.
3. Enter the starting and ending milepoints. If the entire route is being examined, then the default entries of 0 and 1000 can be left as is. If the analyst wants to restrict the search range to, say, the portion of a route within a given region or district, then the appropriate milepoints can be entered.
4. Click "Search" and wait for the results.

The resultant list includes the milepoint of and traffic control at each intersection along the route. Note that it is not possible to distinguish between two-way and four-way stop signs. The analyst can "eliminate" signalized intersections by comparing these lists with those developed using the procedures in sections 7.1 and 7.2. Alternatively, the analyst could copy and paste each resultant list into a spreadsheet. If the analyst adds the crash information at each intersection, then the spreadsheet can be sorted by traffic control type. The shortcoming of the latter approach is that very large databases will be generated (i.e., crash and traffic control data at each intersection along each study route).

7.6 Collision Statistics

As discussed in Section 6, the research team compiled collision-related statistics for a group of 35 study intersections. The statistics compiled included:

• Crash types
• Crash severities
• Direction of travel
• Crash rates
• Collision types
• Functional radius of influence

The retrieval of crash severities is described in section 7.2, while crash rates are described in section 7.4. Crash types and collision types were obtained in a similar manner as crash severities. Upon accessing "Advanced Search" the analyst selects "Accident Type 1" to get information on crash types, and "Collision Type" for collision types. The research team used "Accident Type 1" only, but search fields exist for "Accident Type 2" and "Accident Type 3." The latter types pertain to multiple-vehicle incidents in which more than one crash type was applicable. The research team did not conduct a thorough examination of these additional crash types at the study intersections, but further study is recommended.

The number of crashes by direction of travel at each intersection is a standard output of the CDDS. That is, following each search, the display window summarizes the number of crashes by direction - for state route-state route intersections, all four approaches (legs) are summarized; for state route-federal aid route intersections, only the two legs on the state routes are summarized. The research team observed that the 2004 CDDS summarizes the number of vehicles involved in crashes on each leg. Contrarily, the 2005 CDDS summarizes the number of crashes on each leg. The analyst may find each of these summaries to be useful. To confirm crash totals, however, the 2005 CDDS summary is easier to apply than that of the 2004 CDDS. The research team also observed that the leg assignments at each intersection are straightforward in most applications. At intersections where routes change directions - SR 11 and US 89 in Kanab, for example - the analyst must be careful in assigning legs to routes.

As mentioned previously, the analyst can easily adjust the functional radius of influence on the "Advanced Search" search criteria page, the "Intersections" tool page, or the "SR Intersections" tool page. The CDDS will apply the same radius to all intersections meeting the search criteria. If the analyst wishes to vary the radius, then he or she must run multiple searches on different values. This is a time-consuming exercise, but may be critical to understanding the sphere of influence of an intersection.