Crash Modification Factors (CMFs) are used to estimate the change in crashes due to a change in the roadway. CMFs are reported as a number that is multiplied by the current crash frequency to estimate what the new crash frequency will be once the change is made. CMFs with values greater than 1.0 are expected to INCREASE the crash frequency. CMFs with values less than 1.0 are expected to DECREASE the crash frequency.
Adjustment factors are specialized CMFs that are used in conjunction with the Safety Performance Functions (SPF) of the Highway Safety Manual. With the publication of the HSM in 2010, researchers soon recognized confusion of users concerning the difference between the CMFs used in Part C of the HSM versus the general list of CMFs in Part D. For the 2nd edition of the Highway Safety Manual, the CMFs that are found in Part C will be called Adjustment Factors rather than CMFs. These specialized CMFs are unique because they do not perpetuate the error of applying multiple CMFs to the specific SPFs that referenced them. Adjustment Factors are unique to the SPFs that were used to develop the current models.
There are two commonly used methods for creating CMFs:
Before-After comparison - Comparing the crash frequency (crashes per year) at sites before the countermeasure is implemented, to the crash frequency after the change has been implemented.
Advantages - Can assume a direct causation based on the change.
Disadvantages - Prone to selection bias. Sites are generally selected that have known issues, a true random sample is not obtained. This method takes more time to complete in order to get sufficient data before and after installation.
Cross-Section comparison - Comparing many sites that have one condition with similar sites that have a different condition.
Advantages - Random samples can be better obtained and studies can be done without waiting to collect new data.
Disadvantages - Assumes the study condition is the only difference between the comparison groups. Differences may introduce confounding factors. Furthermore there is no certainty that the condition is directly related to the causation of crashes.
Both study types can implement more stringent statistical methods to help control errors inherent to CMF development. For more information on CMF studies please refer to FHWA’s 2010 document A Guide to Developing Quality Crash Modification Factors.
Some CMFs apply only to specific crash types. These are very useful for analysis, however, they are sensitive to the proportion of crashes being addressed. When applying these CMFs either of the following methods can be used both of which yield the same result:
Option 1: Adjust the crash frequency by the specific crash type. This can be done with Equation 1 as follows:
Napp = Ntotal x P Equation 1
Where:
Napp = Applicable crashes for specific CMF.
Ntotal = Total crashes per year.
P = The proportion of specific crash type to all crashes.
Example:
There are 1.35 crashes per year along a given segment. On the roadway investigated, 37% of total crashes are typically roadway departures (see tables 1-6 of the Historical and Predictive Safety Analysis website for typical proportions of crash types). The CMF selected from the CMF clearinghouse only considered roadway departure crashes when it was developed. Calculate the applicable crashes using Equation 1:
Equation 1: 1.35 x 0.37 = 0.5 applicable crashes. Using the roadway departure CMF of 0.88 results in a reduction of 0.06 crashes: 0.06 = 0.5-(0.5 x 0.88).
Option 2: Adjust the CMF to be applicable to the total crash frequency. This can be done with Equation 2 as follows:
CMFtotal = ( CMFspecific -1) x P +1 Equation 2
Where:
CMFtotal = The CMF that is applicable to the full crash frequency.
CMF specific = The CMF that is applicable to a specific crash type.
P = The proportion of specific crash type to all crashes.
Example:
It is determined that widening the shoulder by 4 feet has a CMF of 0.88 but is only applicable to roadway departure crashes. On a rural two-lane two-way road about 37% of crashes are roadway departure type crashes. What CMF can be applied to all crashes along the roadway?
( 0.88 -1) x 0.37 +1 = 0.956
The CMF applied to the total crash frequency would be 0.956. So if there are 1.35 crashes per year along a given segment and roadway departure type crashes averages 37% of total crashes, the impacts of widening the shoulder can be calculated as 1.35 - (1.35 x 0.956) = 0.06 crashes reduced per year.
The Safe Systems Approach recommends redundancy in our approaches to safety. However, care should be taken when looking at using multiple CMFs at the same time. Many CMFs address similar site conditions and crash types. Applying these overlapping CMFs can suggest greater crash reductions than will be achieved. It is often easier to identify the single lowest CMF and evaluate changes to the crash frequency using this single CMF.
When combining CMFs is necessary, the Investigation of Existing And Alternative Methods for Combining Multiple CMFs (2011) might be used to calculate the total effectiveness of multiple overlapping CMFs. It is important to review the application of the combination methods to ensure that results are reasonable. The FHWA has specific protocols for combining CMFs in their YouTube video, Safety Data and Analysis: Selecting a Method to Analyze Multiple CMFs.
Design elements that are meant to lower the crash risk are referred to as “countermeasures”. Each countermeasure has a CMF associated with it to calculate its effectiveness. UDOT has collected a list of preferred CMFs for common countermeasures to make calculations and comparisons consistent statewide. Be sure to check UDOT’s CMF list regularly for changes that might occur. If you have suggestions for other CMFs feel free to contact crashstudies@utah.gov. Here are some tips help to navigate UDOT Standard CMFs table:
The countermeasures are separated into 3 major types, Intersections, Segments (roadways), and Pedestrians & Other. Each of these major types are provided to break the countermeasures into the various categories. General sub-categories are meant to be used for all intersections or segments. Some of the key column headings are provided below.
Manner of Collision - The crashes that are mitigated by this countermeasure. Countermeasures with “Multiple” suggest that a variety of collisions can be addressed.
Crash Event Addressed - Answers what specific crash types are addressed with the given countermeasure.
Service Years - The estimated service life of the countermeasure. This is used in the Benefit Cost Analysis.
CMF by Crash Severity - These are the recommended CMFs to use by crash severity. 5 = Fatal, 4 = Serious Injury, 3 = Minor Injury, 2 = Possible Injury, 1 = Property Damage Only (no injury).
If you can not find a CMF for your analysis, you might try the CMFclearinghouse.org. This website houses thousands of CMFs based on research efforts from all over the world. When using a CMF from the Clearinghouse be sure to document the CMF #ID for future reference.
When selecting a CMF it is important to review how the CMF was developed (i.e. the research). You should select a CMF where the research is similar to your local conditions. Some key conditions to consider include:
Crash types being addressed.
Roadway and traffic characteristics (i.e. urban/rural, lane count and types, intersection control type, AADT range, etc).
Severity of crashes being addressed.
Reliability of research (e.g. the star rating at the CMF Clearinghouse helps rate the quality of research).
The CMF Clearinghouse often provides several CMFs for the same improvement. Filtering CMFs based on key site information can provide a more useful sample of CMF options. In addition, looking at the general trend in CMF values for the selected group can help you select a better value. Avoid picking extreme values. Be wary of CMFs showing results that are contrary to standard safety practices (e.g. adding lanes to reduce crashes) or that have significantly different results from other studies. These studies are often based on specific problems that are not transferable to every condition.