How We Turn NHTSA Data Into a Verdict

VehicleVerdict flags a model year as a year to avoid when its NHTSA complaint count reaches at least 1.5 times that model's median year, with a minimum of 30 complaints — never by raw complaint count alone — or when an open NHTSA defect investigation is on file. Either condition alone is enough; many flagged years carry both.

Key Takeaways

• A year to avoid is based on complaint rate compared to that model's own median year (1.5x, with a floor of 30 complaints), not a raw complaint total — this keeps a high-volume model like the F-150 from being punished for simply selling more trucks.
• An open NHTSA defect investigation can flag a year on its own, even when complaint volume looks unremarkable.
• NHTSA complaints are self-reported and unverified by NHTSA — VehicleVerdict's numbers measure how often owners reported a problem, not a confirmed defect rate.
• NCAP crash-test stars show up only for years NHTSA actually crash-tested; an untested year shows no rating instead of a guess.
• Years with fewer than 30 complaints get a "limited data" label instead of a verdict, because small samples can swing wildly on a handful of reports.

What counts as a "year to avoid" on VehicleVerdict?

A model year lands on the years-to-avoid list when one of two things is true: its NHTSA complaint count is at least 1.5 times that model's own median year and totals at least 30 complaints, or NHTSA has an open defect investigation tied to that year. Either condition is sufficient on its own. The full rule set is laid out on the methodology page.

We compare each year against its own model's history rather than against a flat number, because a raw complaint count unfairly penalizes vehicles that simply sold well. A model that sold 400,000 units in a given year will naturally generate more complaints than one that sold 40,000, even if it's proportionally the more reliable vehicle. Measuring the complaint rate against the model's own median strips out that volume bias and asks a more useful question: is this particular year unusual for this particular model? The 30-complaint floor exists for the opposite reason — to keep one unlucky year with only a few total complaints from getting flagged off two or three filings that happen to be 1.6 times a tiny median. For more on what's actually inside a single filing, see what is an NHTSA complaint.

What does an open NHTSA investigation add to a verdict?

An open NHTSA defect investigation is one of two standalone triggers for a years-to-avoid flag — it doesn't need a complaint spike to matter. NHTSA opens a Preliminary Evaluation, sometimes upgrades it to an Engineering Analysis, and occasionally issues a Recall Query, all before any recall is announced. A year with an open investigation but an otherwise normal complaint count still gets flagged, because federal investigators are actively looking at a defect pattern that may not have fully shown up in complaint volume yet.

See how PE, EA, and RQ investigations actually work for the difference between those stages. When a year is flagged for both a high complaint rate and an open investigation at the same time, the verdict text calls that out explicitly as two separate red flags rather than one.

How does VehicleVerdict find a model year's dominant problem?

VehicleVerdict looks at how complaints break down by NHTSA's own component categories and calls out a "dominant" problem when one component accounts for at least 40% of all complaints for that year and at least 10 total complaints. That threshold keeps a single loud category in a low-data year from getting labeled "dominant" off a handful of reports.

When no category clears 40% and complaints instead spread across three or more different systems, the page describes that as a diffuse pattern instead — usually a sign of general build-quality variation rather than one specific part failing. Both readings come straight from the component tags NHTSA attaches to each complaint record; what is an NHTSA complaint covers how those records are structured.

What do NHTSA complaints prove — and what don't they?

NHTSA complaints are unverified consumer reports, not confirmed defects. Anyone can file one, and NHTSA does not investigate or validate the underlying claim before logging it. A complaint tells you an owner experienced and reported a problem; it doesn't tell you a mechanic confirmed it, that the part actually failed, or that the failure came from a defect rather than wear, an accident, or misuse.

That's why VehicleVerdict's verdicts describe reporting volume and reporting rate, not a proven defect rate. A model year with a high complaint rate is a year owners complained about more than usual — a real signal worth taking seriously at scale across thousands of filings, but not the same claim as "this many cars actually broke." Check the most-complained model years ranking to see how that volume compares across the whole market.

How do NCAP crash-test ratings factor into a verdict?

NCAP overall star ratings appear on a model-year page only when NHTSA actually crash-tested that specific year — the current five-star format only dates back to 2011, and NHTSA doesn't test every trim and model year of every vehicle. When a rating exists, VehicleVerdict shows it as a closing note on the verdict; when it doesn't, the page says nothing rather than estimating one or carrying over a nearby year's score.

That "never fabricate" rule matters because a redesign or a mid-cycle structural change can shift crash performance from one year to the next, so borrowing a neighboring year's stars would misrepresent a year that was never actually tested. See how NCAP crash-test ratings work for how the stars themselves get calculated.

What are the limits of this approach?

Two limitations matter most. First, small-sample years: a model year with fewer than 30 total complaints doesn't get a years-to-avoid verdict at all — it's labeled limited data instead, because a handful of reports can look statistically extreme without meaning much. Second, self-reporting bias: complaint volume reflects who bothers to file with NHTSA, not a random sample of every problem every owner experienced, so a vocal or online-savvy owner base can generate more complaints than one with an equally real but less-reported issue.

Neither limitation is unique to VehicleVerdict — both are baked into NHTSA's complaint data itself, and the same caveats apply anywhere that data gets used. That's part of why the site leans on complaints, recalls, investigations, and NCAP scores together rather than any single number, and why a years-to-avoid flag should prompt more digging, not serve as a final answer on its own. Researching a used car's NHTSA history before you buy is the natural next step.

The four NHTSA datasets behind every verdict

Every number on a VehicleVerdict page traces back to one of four public NHTSA databases, updated monthly:

1. Consumer complaints — owner-submitted reports filed since 1995, including any injury or death flags and free-text narratives.
2. Safety recalls — official recall campaigns dating back to 1967, each with a defect description, safety consequence, and repair remedy. See how to check for open recalls and how to read a recall notice.
3. Defect investigations — Preliminary Evaluations, Engineering Analyses, and Recall Queries NHTSA opens into suspected defect patterns, whether or not they're still open.
4. NCAP crash-test ratings — overall and category star ratings from physical crash testing, available for tested years from 2011 forward.

VehicleVerdict normalizes vehicle names against NHTSA's own vehicle catalog before combining these four sources, so a "Silverado 1500" and a "Silverado-1500" complaint don't get counted as two different models.

See the years-to-avoid verdict for 12 popular models

This same methodology powers a full years-to-avoid breakdown for every make and model in NHTSA's database — browse them all by make — but these twelve posts cover some of the highest-traffic vehicles on the road, and each shows a different piece of the methodology in action.

Family SUVs and crossovers make up one cluster. The Honda CR-V years to avoid post covers a couple of years where a single powertrain-related component drove the complaint spike. The Toyota RAV4 years to avoid guide shows a year that was flagged mainly by an open investigation rather than complaint volume alone. The Subaru Outback years to avoid breakdown is a clean example of the diffuse-complaint pattern, where no single system dominates. And the Ford Explorer years to avoid page tracks an early redesign year that carried a noticeably heavier complaint load than the generation that followed it.

Trucks and off-roaders form the second cluster. The Ford F-150 years to avoid verdict is a good stress test of the rate-vs-average approach, since it's the best-selling vehicle in the country and would look terrible by raw complaint count alone. The Chevrolet Silverado 1500 years to avoid post shows a generation-change year standing out clearly against the model's own median. The Ram 1500 years to avoid guide is a case where recall counts and complaint rate point in slightly different directions for the same year. The Toyota Tacoma years to avoid page benefits from a long production run, which gives the rate comparison more years of history to work with. And the Jeep Wrangler years to avoid breakdown includes a year where one component made up a clear majority of all filed complaints.

Everyday sedans round out the list. The Toyota Camry years to avoid post needs the rate adjustment just as much as the F-150 does, given how many Camrys are on the road. The Honda Civic years to avoid guide spans several generations, so the model's own median shifts meaningfully from era to era. And the Nissan Altima years to avoid page is a case where severity data — injuries and crashes reported in the complaint narratives — added real weight to a verdict that complaint count alone wouldn't have fully explained.

Frequently Asked Questions

Does a high complaint count always mean a bad model year?

Not by itself. VehicleVerdict compares a year's complaint count to that same model's own median year rather than judging raw numbers, because high-volume models naturally generate more complaints. A model year only gets flagged when its rate is at least 1.5 times the model's median and it has at least 30 total complaints, or when it carries an open NHTSA investigation — a high raw count on a popular model isn't enough by itself.

Can a model year be flagged without any complaint spike at all?

Yes. An open NHTSA defect investigation is enough on its own to put a model year on the years-to-avoid list, regardless of how its complaint count compares to other years. Investigations often start from patterns NHTSA notices before complaint volume catches up, so a year can carry real risk that hasn't fully shown up in the complaint numbers yet.

Does VehicleVerdict verify that each complaint describes a real defect?

No. NHTSA complaints are self-reported by consumers and are not investigated or confirmed by NHTSA before being published, and VehicleVerdict does not independently verify them either. The site's numbers reflect how often a problem was reported, not how many vehicles were confirmed to have failed — a real and useful signal at scale, but not proof of a defect in any single case.

Why do some model years show no NCAP crash-test rating?

Because NHTSA doesn't crash-test every model year of every vehicle, and VehicleVerdict never fills in a missing rating with a guess or a nearby year's score. The current five-star NCAP program only covers vehicles from 2011 forward, and even within that window, testing covers a subset of vehicles each year. If a year was never tested, its page simply omits a rating rather than estimating one.

How current is the underlying NHTSA data?

VehicleVerdict pulls from NHTSA's public complaint, recall, investigation, and NCAP databases, which the agency updates monthly. Each model page reflects the most recent data pull, though very recent model years can still show incomplete complaint histories simply because reporting lags behind a vehicle hitting the road — a newly released model year looks artificially clean for its first several months on sale.

What should I do with a years-to-avoid verdict before buying a used car?

Treat it as a starting point, not a final answer. Look up the specific VIN for open recalls, read a sample of the actual complaint narratives for that year, and check whether the model's flagged issue matches symptoms you can inspect or ask about. The guide to researching a used car's NHTSA history walks through that process step by step.