PE Exam (Civil) Traffic Safety and Accident Analysis: Crash Modification Factors
Last updated: May 2, 2026
Traffic Safety and Accident Analysis: Crash Modification Factors questions are one of the highest-leverage areas to study for the PE Exam (Civil). This guide breaks down the rule, the elements you need to recognize, the named traps that catch most students, and a memory aid that scales to test day. Read it once, then practice the same sub-topic adaptively in the app.
The rule
A Crash Modification Factor (CMF) is the multiplicative ratio of expected crashes WITH a treatment to expected crashes WITHOUT it (AASHTO Highway Safety Manual, Part C). A CMF $< 1$ indicates a safety benefit; a CMF $> 1$ indicates degradation. Predicted crashes follow $N_{predicted} = N_{spf} \cdot (CMF_1 \cdot CMF_2 \cdots CMF_n) \cdot C_x$, where $C_x$ is the local calibration factor. The Crash Reduction Factor is the percentage form: $CRF = (1 - CMF) \times 100\%$.
Elements breakdown
CMF Definition
Ratio capturing the safety effectiveness of a single countermeasure relative to a baseline condition.
- Identify base condition the CMF was developed for
- Confirm crash type and severity scope
- Confirm facility type matches your site
- Read CMF value from HSM or CMF Clearinghouse
- Note standard error and star rating
CMF vs CRF Conversion
Translate between the multiplier (CMF) and the percent-reduction form (CRF) without flipping signs.
- $CRF = 1 - CMF$ when CMF $< 1$
- $CRF$ is negative if CMF $> 1$ (treatment increases crashes)
- Express CRF as percent for reporting
- Apply CMF (not CRF) directly to crash counts
- Never subtract CRF from a count
Combining Multiple CMFs
Apply several independent treatments simultaneously using the multiplicative assumption.
- Verify treatments act on independent crash mechanisms
- Multiply CMFs: $CMF_{combined} = \prod CMF_i$
- Avoid double-counting overlapping countermeasures
- Apply engineering judgment when more than three CMFs combine
- Cap reductions when combined CMF unrealistically low
HSM Predictive Method
Estimate site-specific crash frequency using a Safety Performance Function adjusted by CMFs and calibration.
- Select correct SPF for facility type
- Compute baseline $N_{spf}$ from AADT and length
- Apply each site-condition CMF as multiplier
- Multiply by local calibration factor $C_x$
- Optionally combine with observed counts via Empirical Bayes
CMF Quality Screening
Use the FHWA CMF Clearinghouse star rating to gauge confidence before applying a value.
- 5-star: highest reliability, large sample, controlled study
- 3-star: moderate reliability, acceptable for screening
- 1- to 2-star: directional only, not for design selection
- Match study setting to your project context
- Document standard error in the design memo
Benefit Calculation
Convert crash reduction into monetary safety benefit for project prioritization.
- Compute crashes prevented: $\Delta N = N_{baseline}(1 - CMF)$
- Multiply by comprehensive cost per crash
- Sum severity-weighted benefits if CMF is severity-specific
- Compare to service-life cost for $B/C$ ratio
- Use FHWA or state DOT crash unit costs
Common patterns and traps
The Additive Combination Trap
When two or more treatments are proposed, the wrong instinct is to sum CRFs (e.g., $14\% + 7\% = 21\%$). The HSM rule is multiplicative: $CMF_{combined} = \prod CMF_i$. Adding overstates benefits as more CMFs stack and can even drive total reduction above 100\%.
A choice equal to $N_{obs} \times (1 - \sum CRF_i)$ rather than $N_{obs} \times \prod CMF_i$.
The CMF/CRF Mix-Up
The CMF is the post-treatment survivor fraction; the CRF is the percent reduction. A candidate who multiplies the count by the CRF gets the eliminated portion, not the remainder. The error is sign-like: it inverts the answer.
A choice equal to $N_{obs} \times CRF$ instead of $N_{obs} \times CMF$, often the smallest answer in the set.
The Calibration Factor Drop
The HSM predictive method ends with multiplication by a jurisdiction-specific calibration factor $C_x$ that adjusts national SPFs to local conditions. Forgetting $C_x$ produces an answer that looks plausible but is systematically biased low when $C_x > 1$.
A choice equal to $N_{spf} \cdot \prod CMF_i$ with no $C_x$ multiplier.
The Wrong Crash-Type Application
CMFs are scoped to a specific crash type and severity (total, fatal+injury, run-off-road, head-on, etc.). Applying a CMF developed for fatal+injury crashes to a total crash count overstates or understates the benefit.
A choice that applies a fatal-only CMF (often a more aggressive number like $0.55$) to a total-crash baseline.
The Low-Star CMF Overconfidence
The FHWA CMF Clearinghouse rates CMFs from 1 to 5 stars. A 1- or 2-star CMF carries large standard error and should not drive design selection or B/C ranking. Treating a 1-star CMF as exact yields a precise-looking but unreliable benefit estimate.
A choice computed with a 1-star CMF reported to four significant figures, ignoring stated $\pm$ standard error bounds.
How it works
Picture a 4-mile rural two-lane segment averaging $N_{obs} = 20$ crashes per year. You propose installing centerline rumble strips with $CMF = 0.86$ for total crashes (HSM Chapter 13). The expected post-treatment frequency is $N_{after} = 20 \times 0.86 = 17.2 \text{ crashes/yr}$, a reduction of $\Delta N = 20 - 17.2 = 2.8 \text{ crashes/yr}$, or $CRF = (1 - 0.86) \times 100\% = 14\%$. Now suppose you also widen shoulders ($CMF = 0.93$, independent crash mechanism). The combined CMF is $0.86 \times 0.93 = 0.7998 \approx 0.80$, so $N_{after} = 20 \times 0.80 = 16.0 \text{ crashes/yr}$. Notice how adding the CRFs ($14\% + 7\% = 21\%$) gives a different and incorrect answer of $20 \times (1 - 0.21) = 15.8$. The arithmetic gap is small here, but PE distractors are engineered to land exactly on this kind of additive shortcut.
Worked examples
You are evaluating a $4.0 \text{-mi}$ segment of two-lane rural highway on the Reyes Bypass. The observed crash history over the past five years averages $N_{obs} = 18 \text{ crashes/yr}$ for total crashes. The county is considering installing centerline rumble strips along the full segment. The applicable HSM CMF for centerline rumble strips on rural two-lane highways is $CMF = 0.86$ for total crashes, taken from a 4-star Clearinghouse entry. No other treatments are proposed. Assume traffic volumes and roadway geometry remain unchanged after construction.
Most nearly, what is the expected total crash frequency after the treatment is installed?
- A $2.5 \text{ crashes/yr}$
- B $15.5 \text{ crashes/yr}$ ✓ Correct
- C $17.1 \text{ crashes/yr}$
- D $20.9 \text{ crashes/yr}$
Why B is correct: Apply the CMF directly to the observed frequency: $N_{after} = N_{obs} \cdot CMF = 18 \times 0.86 = 15.48 \text{ crashes/yr}$, which rounds to $15.5 \text{ crashes/yr}$. The unit check is $\text{crashes/yr} \times \text{(dimensionless)} = \text{crashes/yr}$, confirming the answer. Choice B matches.
Why each wrong choice fails:
- A: This applies the CRF directly: $18 \times (1 - 0.86) = 18 \times 0.14 = 2.52$. That gives the crashes prevented, not the crashes remaining. (The CMF/CRF Mix-Up)
- C: This subtracts the CMF as if it were a count: $18 - 0.86 \approx 17.1$. Subtraction has no dimensional basis because the CMF is a unitless multiplier.
- D: This inverts the multiplier and uses $18 \times (1 + 0.14) = 20.5$ (or similar), treating the CMF as if it increased crashes. Direction matters: $CMF < 1$ always reduces. (The CMF/CRF Mix-Up)
A four-leg unsignalized intersection on the Liu Civic Center access road averages $N_{obs} = 12 \text{ crashes/yr}$ for total crashes. The city is bundling two independent treatments: (1) convert the minor-street stop control to all-way stop control, with $CMF_1 = 0.52$ for total crashes, and (2) add intersection lighting, with $CMF_2 = 0.78$ for total crashes (HSM Chapter 14, both 4-star Clearinghouse entries). The treatments act on independent crash mechanisms, so the multiplicative assumption is appropriate. No traffic-volume change is anticipated.
Most nearly, what is the combined Crash Reduction Factor (CRF) for the bundled treatment?
- A $35\%$
- B $41\%$
- C $59\%$ ✓ Correct
- D $70\%$
Why C is correct: Combine CMFs multiplicatively: $CMF_{combined} = CMF_1 \cdot CMF_2 = 0.52 \times 0.78 = 0.4056$. Convert to CRF: $CRF = (1 - 0.4056) \times 100\% = 59.4\%$, which rounds to $59\%$. The check: applying to the count gives $N_{after} = 12 \times 0.4056 = 4.87 \text{ crashes/yr}$, a reduction of about $7.1 \text{ crashes/yr}$, consistent with $\sim 59\%$ of $12$.
Why each wrong choice fails:
- A: This averages the two CMFs: $\frac{0.52 + 0.78}{2} = 0.65$, then $1 - 0.65 = 35\%$. CMFs are not averaged; they are multiplied. (The Additive Combination Trap)
- B: This reports the combined CMF as a percent ($0.4056 \approx 41\%$) instead of converting to CRF. The candidate forgot the $1 - CMF$ step. (The CMF/CRF Mix-Up)
- D: This adds the individual CRFs: $(1-0.52) + (1-0.78) = 0.48 + 0.22 = 0.70 = 70\%$. Adding CRFs systematically overstates the combined benefit. (The Additive Combination Trap)
You are applying the HSM predictive method to a $2.5 \text{-mi}$ rural two-lane highway segment on the proposed Vasquez Ranch Road realignment. The Safety Performance Function returns a base predicted frequency of $N_{spf} = 14.0 \text{ crashes/yr}$ for the base conditions. The site differs from base conditions in two ways: lane width is $11 \text{ ft}$ ($CMF_{lw} = 1.05$) and shoulder width is $4 \text{ ft}$ ($CMF_{sw} = 1.10$). The state DOT has developed a local calibration factor $C_x = 1.15$ for this facility type. Assume the two CMFs are independent.
Most nearly, what is the predicted total crash frequency for the segment?
- A $16.1 \text{ crashes/yr}$
- B $16.2 \text{ crashes/yr}$
- C $18.6 \text{ crashes/yr}$ ✓ Correct
- D $46.2 \text{ crashes/yr}$
Why C is correct: Apply the HSM predictive equation: $N_{predicted} = N_{spf} \cdot CMF_{lw} \cdot CMF_{sw} \cdot C_x = 14.0 \times 1.05 \times 1.10 \times 1.15$. Compute stepwise: $14.0 \times 1.05 = 14.70$; $14.70 \times 1.10 = 16.17$; $16.17 \times 1.15 = 18.60 \text{ crashes/yr}$. Units: $\text{crashes/yr} \times \text{(dimensionless)}^3 = \text{crashes/yr}$.
Why each wrong choice fails:
- A: This applies only the calibration factor: $14.0 \times 1.15 = 16.1$. The candidate dropped both site-condition CMFs.
- B: This applies the two CMFs but omits $C_x$: $14.0 \times 1.05 \times 1.10 = 16.17$. The local calibration step was skipped. (The Calibration Factor Drop)
- D: This sums the multipliers: $14.0 \times (1.05 + 1.10 + 1.15) = 14.0 \times 3.30 = 46.2$. CMFs and $C_x$ are combined multiplicatively, not additively. (The Additive Combination Trap)
Memory aid
"Multiply, don't add; CMF for counts, CRF for percent." Then ask: did I include $C_x$?
Key distinction
The CMF is the survivor fraction (what remains after treatment); the CRF is the prevented fraction (what was eliminated). Multiply CMFs by counts and by each other; convert to CRF only for reporting reductions in percent.
Summary
CMFs are multiplicative survivor fractions: chain them together with the SPF and the local calibration factor, and only convert to CRF when you need a percent.
Practice traffic safety and accident analysis: crash modification factors adaptively
Reading the rule is the start. Working PE Exam (Civil)-format questions on this sub-topic with adaptive selection, watching your mastery score climb in real time, and seeing the items you missed return on a spaced-repetition schedule — that's where score lift actually happens. Free for seven days. No credit card required.
Start your free 7-day trialFrequently asked questions
What is traffic safety and accident analysis: crash modification factors on the PE Exam (Civil)?
A Crash Modification Factor (CMF) is the multiplicative ratio of expected crashes WITH a treatment to expected crashes WITHOUT it (AASHTO Highway Safety Manual, Part C). A CMF $< 1$ indicates a safety benefit; a CMF $> 1$ indicates degradation. Predicted crashes follow $N_{predicted} = N_{spf} \cdot (CMF_1 \cdot CMF_2 \cdots CMF_n) \cdot C_x$, where $C_x$ is the local calibration factor. The Crash Reduction Factor is the percentage form: $CRF = (1 - CMF) \times 100\%$.
How do I practice traffic safety and accident analysis: crash modification factors questions?
The fastest way to improve on traffic safety and accident analysis: crash modification factors is targeted, adaptive practice — working questions that focus on your specific weak spots within this sub-topic, getting immediate feedback, and revisiting items you missed on a spaced-repetition schedule. Neureto's adaptive engine does this automatically across the PE Exam (Civil); start a free 7-day trial to see your sub-topic mastery climb in real time.
What's the most important distinction to remember for traffic safety and accident analysis: crash modification factors?
The CMF is the survivor fraction (what remains after treatment); the CRF is the prevented fraction (what was eliminated). Multiply CMFs by counts and by each other; convert to CRF only for reporting reductions in percent.
Is there a memory aid for traffic safety and accident analysis: crash modification factors questions?
"Multiply, don't add; CMF for counts, CRF for percent." Then ask: did I include $C_x$?
What's a common trap on traffic safety and accident analysis: crash modification factors questions?
Adding CRFs instead of multiplying CMFs
What's a common trap on traffic safety and accident analysis: crash modification factors questions?
Confusing CMF with CRF (using 0.14 where 0.86 belongs)
Ready to drill these patterns?
Take a free PE Exam (Civil) assessment — about 35 minutes and Neureto will route more traffic safety and accident analysis: crash modification factors questions your way until your sub-topic mastery score reflects real improvement, not luck. Free for seven days. No credit card required.
Start your free 7-day trial