Which BASIC Scores Actually Predict a Crash — and Which Are Just Noise

A carrier at the 85th percentile in Cargo-Related BASIC gets flagged the same shade of red as one at the 85th percentile in Unsafe Driving. Most screening tools treat those two alerts identically. I've watched brokers decline carriers because of a high Cargo-Related score on a dry van load — without stopping to consider that the metric barely moves the crash-risk needle for carriers that don't haul flatbed or open-deck.

Seven BASICs. Seven different stories. And they're not equal.

Here's how I actually weight them, ranked by how much they tell me about the probability that a carrier is going to be involved in a serious crash in the next twelve months.

Tier One: These Actually Predict Crashes

Unsafe Driving is the one that matters most. FMCSA's own research shows the Unsafe Driving BASIC has the strongest correlation with future crash involvement of any category. It captures roadside inspection violations tied to speeding, reckless driving, improper lane changes, and failure to obey traffic control devices — the conduct covered under 49 CFR § 392.2 (obedience to state and local traffic laws) and § 392.3 (ill or fatigued operators). These are the behaviors that show up in crash reconstructions.

When a carrier is at or above the 65th percentile in Unsafe Driving, I want to see the underlying violations before I make a decision — not just the percentile. Was it three written warnings for 5 mph over on a flat interstate in 2022? Or was it a citation for 22 over on a mountain grade six months ago? The percentile compresses that distinction into a single number. The violation list doesn't.

Vehicle Maintenance is the other strong predictor. Brake violations dominate this BASIC — 49 CFR § 393.47 (brake performance) and § 393.48 (brakes required to be operative) show up constantly. Brake-related defects appear in roughly 44% of large-truck crash causes in FMCSA's Large Truck Crash Causation Study. A carrier with a high Vehicle Maintenance score is telling you their shop discipline is weak. That doesn't mean they'll crash tomorrow. It means the underlying condition that produces crashes is present and not being managed.

For both of these, I run a full screening on every carrier above the 65th percentile and look at violations grouped by inspection date. A cluster from one bad inspection event is different from a steady drip across twenty inspections over eighteen months. The pattern tells you whether this is an isolated bad day or a systemic problem.

Tier Two: Real Signal, Complex Interpretation

Crash Indicator sounds like the most obvious predictor of future crashes. It's actually murkier than Unsafe Driving or Vehicle Maintenance for one reason: preventability isn't built in. The Crash Indicator BASIC counts DOT-reportable crashes — any incident involving a fatality, an injury requiring immediate medical treatment, or disabling damage requiring a tow — but it doesn't automatically assign fault. A carrier can accumulate a high Crash Indicator score from a string of rear-end strikes against their trucks at loading docks by inattentive four-wheelers.

That doesn't mean you ignore it. A high Crash Indicator score with zero DataQ challenges is a yellow flag. If the carrier hasn't contested any crashes, they either don't know they can, or they know the crashes were their fault. Either is worth a phone call to the safety director.

A high score with multiple DataQ challenges — especially successful removals — is a different profile. Do the work before you decide.

Hours of Service percentile is real signal that requires context. The HOS BASIC captures violations under 49 CFR § 395.3 (maximum driving time for property-carrying vehicles) and § 395.8 (driver's record of duty status). High HOS scores can mean genuine hours violations — drivers running past their 11-hour limit, or falsifying logs. They can also mean a high inspection frequency at a terminal that happens to be on a targeted enforcement rotation.

The ELD mandate changes how you read this BASIC. Pre-ELD, HOS violations were easier to explain away because paper logs were easy to falsify and violations were caught inconsistently. Post-ELD, a violation you see in the BASIC is harder to dispute. A carrier with a high HOS score from the last 24 months, running on ELDs the whole time, doesn't get the benefit of the doubt that a violation cluster from 2018 might.

Driver Fitness measures violations under 49 CFR § 391.11 (general qualifications of drivers) and § 391.41 (physical qualifications). That covers operating without a valid CDL, driving with a suspended or revoked license, and medical certificate violations. It's not the highest-frequency BASIC, but when it's elevated it tells you something specific: this carrier doesn't have tight control over driver qualification management. That failure tends to cluster. A carrier who misses CDL status checks also misses medical cert expirations. It's the same underlying problem.

Tier Three: Low Frequency, High Severity

Controlled Substances/Alcohol is exactly what it says. 49 CFR Part 382 requires pre-employment testing, post-accident testing, random testing at a minimum 50% annual rate for CDL drivers, and return-to-duty testing after a violation. A carrier with any violations in this BASIC in the last 24 months gets a phone call before I tender a load. I don't care if the percentile is at the 40th. Any violation is a management failure worth exploring.

The reason this lands in tier three isn't importance — it's base rate. Most carriers don't show violations here. When one does, the severity potential is high enough that even a single data point deserves scrutiny, independent of where the percentile ranks.

A carrier at MC-1247893 / DOT-3567102 I screened earlier this year had a 41st percentile CS score — barely flagged by most tools. The underlying data showed a single failed pre-employment test from 14 months ago that didn't appear to have been followed by a completed return-to-duty protocol in FMCSA's system. I called them. They had no record of this driver ever completing a SAP evaluation. I didn't tender the load.

Tier Four: Context-Dependent, Often Noise

Cargo-Related captures violations under 49 CFR Part 393 Subpart I (protection against shifting or falling cargo) and § 393.104 (tie-down assemblies). High Cargo-Related scores are predictive for flatbed and open-deck carriers. For dry van or reefer carriers, they're largely noise.

A dry van carrier doesn't have exposed cargo that roadside inspectors examine from the outside. Their Cargo-Related violations, when they exist, typically come from door seal issues or the occasional partial inspection where interior cargo wasn't secured properly. That's not the same risk profile as a flatbed carrier with a pattern of improperly secured steel coils.

This is where I see the most mistaken declinations. A broker sees an 82nd percentile Cargo-Related score and passes on a carrier. The carrier moves temperature-controlled pharmaceuticals under controlled conditions. Their Unsafe Driving BASIC is at the 18th percentile. Vehicle Maintenance is at the 22nd. The Cargo-Related score is from two inspection events on a specialized haul they no longer run.

When you see an elevated Cargo-Related score, look at the equipment type in the underlying inspections before you draw a conclusion.

Hazmat applies to carriers moving regulated quantities of hazardous materials under 49 CFR Part 397 and the broader HMR requirements at 49 CFR Parts 171–180. For carriers that don't haul hazmat, this BASIC is functionally irrelevant. Even for carriers that do, the inspection volume for hazmat-specific checks is low enough that percentiles are sensitive to small absolute numbers. One placard violation in two years can put a carrier at a surprisingly high percentile relative to peers.

Don't use the Hazmat percentile as a proxy for general safety on a general freight carrier. It isn't one.

Putting It Together

A carrier at the 80th percentile in Unsafe Driving AND the 80th percentile in Vehicle Maintenance is a fundamentally different risk profile than one at the 90th percentile in Cargo-Related with everything else clean. The first combination has two of the highest-predictive BASICs elevated simultaneously — that's a pattern that shows up repeatedly in carrier files that end up in post-crash discovery. The second might be a flatbed carrier with inconsistent securement discipline, or it might be noise from two unusual inspection events.

The combined read I use: if Unsafe Driving is above the 65th percentile, I want violation details before I decide. If Vehicle Maintenance is above the 65th percentile, same. If both are above 65th, the combined signal is stronger than either alone — carriers don't typically score high in both without systemic management failures. Everything else I evaluate in context of what the carrier actually hauls, the equipment type in the underlying violations, and the direction of the trend over the last 24 months.

That last piece matters. A carrier who was at the 78th percentile in Vehicle Maintenance eighteen months ago and is now at the 52nd — with a recent fleet replacement and a new shop director — is telling a different story than a carrier whose score has been climbing steadily for two years.

After Montgomery v. Caribe Transport II, LLC, the question in discovery isn't just "did you check the BASICs?" It's "did you understand what you were looking at?" Checking a box against a threshold without understanding what the scores mean in context of the equipment, the cargo type, and the violation pattern is a thin paper trail if a serious crash follows a tender.

How I Document This

When DOTScreener flags elevated BASIC scores at screening, I add a note to the carrier file before the load tenders. It includes:

• Which BASICs were elevated and the percentile at time of screening
• What violation types drove the elevation, pulled from the underlying inspection data
• The reasoning for proceeding or not: equipment match to cargo type, trend direction over the last 24 months, whether the violations are concentrated or spread
• The date and time of the pull

If I chose to proceed with a carrier who had an elevated Unsafe Driving or Vehicle Maintenance score, that reasoning is explicit in the file. Not because I expect to be sued. Because the reasoning that was sound at load-tender time is the only thing that holds up later if the question gets asked.

---

— Mason Lavallet

Founder, DOTScreener.com