Most brokers glance at BASIC scores. Plaintiffs' lawyers read them carefully. Here's what each category actually predicts, what the 65% and 80% thresholds mean, and which BASIC tends to be the most dangerous when ignored.

The first time I sat through a deposition transcript from a broker-liability case, I expected the plaintiff's lawyer to grill the broker on the obvious stuff — was the authority active, was insurance in force, was the carrier on a denied-list. He spent maybe four minutes on those. He spent the next ninety minutes on the BASIC categories. By the end the broker's safety director was conceding things on the record that I'm pretty sure he wished he hadn't.

That experience reorganized how I think about Compliance, Safety, Accountability (CSA) scores. Brokers tend to look at them as a single number — "are they alert anywhere?" Plaintiffs' lawyers read them as seven separate stories. Each category is a different argument. If you know which arguments they're building, you can decide ahead of time whether to take the load.

What the BASIC categories actually are

FMCSA's Safety Measurement System breaks carrier safety into seven Behavior Analysis and Safety Improvement Categories. Six of them have public percentile scores; one (Driver Fitness on the public site) only shows alert status. They are:

Unsafe Driving — speeding, lane changes, reckless driving, failure to use seat belts. Inspection-derived from roadside.
Hours-of-Service (HOS) Compliance — log violations, driving over hours, failure to maintain ELD records. Also inspection-derived.
Driver Fitness — driver qualification violations: medical card, CDL, training. Inspection plus investigation.
Controlled Substances/Alcohol — drug and alcohol violations.
Vehicle Maintenance — brakes, lights, tires, cargo securement failures observed at inspection.
Hazardous Materials Compliance — applies only if the carrier is a hazmat carrier.
Crash Indicator — DOT-reportable crashes in the last 24 months. *Not* fault-adjusted on the public site — every recordable crash counts, preventable or not.

A carrier's percentile in each category is computed against a peer group of carriers of similar size and operation. A 75 in Unsafe Driving means the carrier scored worse than 75% of similar carriers.

The thresholds — and why they're not the same number everywhere

This is where most brokers get tripped up. FMCSA does not use a single percentile threshold. It uses three sets, depending on the carrier:

General carriers: 65% for Unsafe Driving, HOS Compliance, and the Crash Indicator. 80% for Driver Fitness, Controlled Substances/Alcohol, and Vehicle Maintenance.

Hazmat (placardable) carriers: 60% for Unsafe Driving, HOS Compliance, and Crash Indicator. 75% for the others.

Passenger carriers: 50% for Unsafe Driving, HOS Compliance, and Crash Indicator. 65% for the others.

The reason is exactly what you'd expect — bus crashes and hazmat releases have higher consequences, so the agency catches those carriers earlier in the percentile curve. For broker vetting purposes, the practical takeaway is that **65% in Unsafe Driving on a general-freight carrier is the line beyond which FMCSA itself has flagged the carrier for elevated intervention**. If the agency has already flagged them, ignoring it in your file is a hard sell in deposition.

Which BASIC is the most dangerous to ignore

Pop quiz I ask brokers when I'm training them: of the seven BASICs, which one shows up most often in fatality litigation against the broker?

Most guess Crash Indicator. Some guess Vehicle Maintenance. The right answer is almost always **Unsafe Driving**.

Here's why. The Crash Indicator counts crashes that already happened. The plaintiff's lawyer doesn't need the score to prove a crash happened — they have the police report. The Crash Indicator is a useful tally, but it doesn't add narrative force.

Unsafe Driving is different. Unsafe Driving is a record of *behavior leading up to the crash you're now litigating*. Speeding tickets. Lane-change violations. Following-too-close citations. The plaintiff's lawyer puts the BASIC printout on the easel, points to the 78% percentile flagged in Unsafe Driving four months before the wreck, and asks the broker rep: "Did you look at this before you tendered the load?"

If the answer is no — and the carrier hit the plaintiff at speed — that one BASIC just told a jury that the broker dispatched a load to a carrier whose drivers were already on record for speeding more than 78% of their peers. That is the negligent-selection argument in one slide.

A concrete scenario

You're vetting a reefer carrier for a Texas-to-California run. MC-1338207 / DOT-2841650. Five power units, two years of authority. Insurance L&I clean. Crash Indicator at 22 — below threshold, fine. Vehicle Maintenance at 49 — below threshold, fine. HOS at 31 — fine.

Unsafe Driving at **74**. Alert.

What does that 74 mean concretely? It means in the last 24 months of roadside inspections, this carrier's drivers were cited for speeding, lane changes, following-too-close, or similar offenses at a rate worse than 74% of peer carriers their size. Specific violations. Real records. Already in the public file.

Do you take the load? I might, depending on the trend and the load profile. If the percentile is dropping month over month and the carrier can explain — they fired a driver last year, they implemented an in-cab camera program in February, they can show the cite-rate per inspection has come down — that's a real conversation. If the percentile has been north of 65 for the last six months and nobody at the carrier can explain why, I'm passing.

Either way, **whatever I decide, the decision and the reason for it go in the file**. After *Montgomery v. Caribe Transport II*, the question isn't whether I had to refuse the carrier. The question is whether I made a documented, reasoned decision.

The trap: a carrier with no scores

FMCSA's CSA SMS uses inspection events to compute percentiles. A carrier with very few inspections won't have a public percentile in some categories — the data sample is too small. The SMS will show "Insufficient Data" or no public alert.

Lots of brokers read this as "no problem." It is not. A two-year-old carrier with three power units and zero roadside inspections is not a carrier with great safety performance. It's a carrier no inspector has had eyes on in two years. The CSA percentiles aren't useful here — but neither is the absence of them.

For carriers in this bucket, the diligence record has to lean harder on the other signals: authority age, MCS-150 last-update, insurance L&I (real or stale?), and the attestation flow. Pre-Montgomery a lot of brokers used "no negative BASIC scores" as a green light. Post-Montgomery that's a coin flip you're betting your file on.

The regulation, in plain English

CSA isn't a regulation that creates a private right of action — but the underlying violations are. 49 CFR Part 392 (driving rules), Part 395 (HOS / ELD), Part 391 (driver qualification), Part 393 (vehicle inspection, repair, and maintenance), Part 396 (inspection/repair/maintenance generally) are the source statutes. Roadside violations against these parts feed into BASIC. So when you read "Unsafe Driving 74%," what you're really reading is "this carrier's drivers have been getting cited under Part 392 at a rate worse than 74% of their peer group."

That is admissible. That is what gets pointed to during the broker's deposition.

How I document this

Three artifacts per carrier at lookup time:

1. **A timestamped snapshot of all BASIC percentiles** as displayed in SMS. Not just the alerts — all of them, including the "Insufficient Data" categories. Discovery requests sweep wider than people think.

2. **A note on any BASIC at or near threshold**, with a short written rationale for the tender decision. One paragraph. "Unsafe Driving 74% — confirmed with carrier safety director on 5/12 that two drivers cited in Q4 are no longer with the carrier; ELD rollout 2/2026. Approved with mitigation: pre-trip attestation required, no team driver assignment." That's enough. The point is showing thought.

3. **The peer-group context**. SMS shows what the carrier's threshold is for their type (general/hazmat/passenger). If you screen a passenger carrier you should know their threshold is 50%, not 65%. Including this in the file shows you understood the framework.

You're not trying to prove the carrier was perfect. You're proving you looked, you understood what you were looking at, and you made a reasoned call. That's reasonable diligence.

— Mason Lavallet

Founder, DOTScreener.com

---

Sources

[FMCSA Safety Measurement System Methodology (PDF)](https://csa.fmcsa.dot.gov/Documents/SMSMethodology.pdf) — the official scoring methodology
[FMCSA SMS Tools](https://ai.fmcsa.dot.gov/SMS/Tools/Index.aspx) — public carrier SMS lookups
[FMCSA — Compliance, Safety, Accountability (CSA) Overview](https://csa.fmcsa.dot.gov/about/Measure)
[Foley Services — FMCSA Tweaks Intervention Thresholds for Certain Carriers](https://www.foleyservices.com/news/fmcsa-tweaks-intervention-thresholds-for-certain-carriers/) — threshold by carrier type
[Hanson Bridgett — Montgomery v. Caribe Transport II analysis](https://www.hansonbridgett.com/publication/260514_8509_supreme-court-faaaa)

Reading FMCSA BASIC Percentiles Like a Plaintiff's Lawyer