PULSCAR ANALYSIS OF NHTSA PUBLIC DATA — WHEEL BEARINGS AND HUB ASSEMBLIES PulsCar analysis of NHTSA public data, September 2026. License: CC BY 4.0. Files in this folder may be reused with attribution to PulsCar (pulscar.io). -------------------------------------------------------------------------- 1. SOURCE FILES -------------------------------------------------------------------------- NHTSA public flat files downloaded from static.nhtsa.gov and loaded into a local SQLite database: Complaints FLAT_CMPL.txt 777,040 owner complaints (ODINO), received 1 January 2015 - 8 September 2026. The set below runs 6 January 2015 - 26 August 2026. Field definitions: https://static.nhtsa.gov/odi/ffdd/cmpl/CMPL.txt Recalls FLAT_RCL.txt one row per make/model/model year per campaign; campaigns counted DISTINCT on the campaign number. Communications FLAT_TSBS.txt manufacturer communications; counted DISTINCT on the NHTSA document id. Complaints are owner reports: not verified repairs, not failure rates, and with no denominator of vehicles on the road. A make or model with more complaints may simply have sold more cars, or older ones. Nothing here ranks reliability. -------------------------------------------------------------------------- 2. THE COMPLAINT QUERY (1,762 -> 1,369 -> 1,270) -------------------------------------------------------------------------- Step 1 - full-text search (FTS5) for any of: "wheel bearing", "wheel bearings", "hub bearing", "bearing hub", "hub assembly", "wheel hub" -> 1,762 complaints. Step 2 - failure proximity. The phrase must sit within 80 characters of a failure, noise or replacement word: fail* | bad | worn | wear | noise | noisy | hum* | growl* | grind* | roar* | replac* | seiz* | separat* | came off | fell off | broke* | collaps* | play | loose | defect* | went out | gone bad | rumbl* | whin* | damag* -> 1,369 complaints. Step 3 - passenger vehicles only. 99 complaints filed under travel-trailer, motorhome, boat-trailer and motorcycle makes (Grand Design, Harley-Davidson, Forest River, Karavan, Jayco, Keystone, Winnebago, Kawasaki and others) were removed, because a trailer wheel bearing is a different part with a different service life and would distort the mileage figures. They remain visible in the recall file and in this note. -> 1,270 complaints. Every figure on the page uses this set. Step 4 - SAMPLE READING. 30 random members of the pre-exclusion 1,369 were read in full (random.seed(9)). 24 were genuine passenger-vehicle wheel bearing or hub reports, 4 were the trailer and motorcycle reports that step 3 then removed, and 2 were cases where a bearing was named but was not the fault: one where the noise turned out to be a power transfer unit whose seal had failed, and one where an owner speculated about bearing wear without a diagnosis. Observed false-positive rate after step 3: about 7% (2/30). -------------------------------------------------------------------------- 3. WHAT IS COUNTED (bearing-complaints-by-context.csv) -------------------------------------------------------------------------- Category searches run over the whole narrative, not a proximity window. A complaint can appear in several rows; shares are of 1,270, not parts of a whole. The towed / crash / fire columns are NHTSA's own structured flags, published next to the all-file baselines over all 777,040 complaints: towed 5.11%, crash 4.37%, fire 1.82%. Mileage: values <= 0 and >= 400,000 dropped; 853 of 1,270 carry a usable figure. Median 76,500 miles, Q1 46,888, Q3 113,000. 142 of the 853 (16.6%) report the failure at 36,000 miles or less and 308 (36.1%) below 60,000. That is the middle of a pile of REPORTED failures, not a service life: owners file when a part fails early, so this distribution is biased toward early failure by construction. It is quoted on the page only to make the warranty point, never as a lifespan. "The wheel came off, separated or nearly did" (28 complaints) was read individually; the phrasing ranges from a wheel that detached while driving to one a shop found about to. It is published as a count, not a rate. -------------------------------------------------------------------------- 4. MODELS (bearing-complaints-top-models.csv) -------------------------------------------------------------------------- Make/model pairs with 6 or more complaints, with the three most common model years, median reported mileage, how many narratives name a rear bearing, how many name an ABS or traction control light, and how many report the failure at 36,000 miles or less. Complaint records do not identify the trim, the axle or the bearing design, so nothing here can tell you whether a given car uses a bolt-on hub or a pressed-in bearing. -------------------------------------------------------------------------- 5. RECALLS (bearing-and-hub-recalls.csv) -------------------------------------------------------------------------- Campaigns whose component, defect text or remedy text names a wheel bearing, hub bearing, hub assembly or wheel hub: -> 131 distinct campaigns, 726,868 vehicles and pieces of equipment potentially affected (sum of per-campaign figures; a vehicle covered twice is counted twice). Two subsets quoted on the page, both computed from this file: 23 campaigns are EQUIPMENT recalls of replacement parts rather than vehicles (123,017 units), flagged in the campaign_type column, and 28 campaigns describe a wheel that may separate or detach (70,183 units). Trailers, motorhomes and heavy trucks are included and visible in the make column. -------------------------------------------------------------------------- 6. LIMITATIONS -------------------------------------------------------------------------- - Self-selected reports; filers are not a random sample of owners. - No denominator: model counts track sales volume and fleet age. - Regexes read words, not meaning; negation is counted as a match. - Complaints rarely state what finally fixed the car, so this analysis counts what was NAMED, not what was proven. In particular it cannot measure how often a "bearing" turned out to be tires. - Nothing here measures how long a bearing lasts, and the mileage distribution must not be read as a service life. Files: bearing-complaints-by-context.csv, bearing-complaints-by-make.csv, bearing-complaints-top-models.csv, bearing-and-hub-recalls.csv. ------------------------------------------------------------------------------ CASE SENSITIVITY — SECTION ADDED 19 SEPTEMBER 2026 ------------------------------------------------------------------------------ NHTSA stores complaint narratives in ALL CAPS for records filed through roughly 2020 and in mixed case for records filed from 2021 on. A case-SENSITIVE search of the narrative field loses about 92% of the complaints filed after 2020, so this is the single largest way an analysis of this database can go wrong while still looking finished. Every query behind this file is case-insensitive. SQLite's LIKE and its FTS5 full-text index both ignore case by default, and no query in this work uses GLOB or any case-sensitive comparison. This was not taken on trust. On 19 September 2026 every published folder was re-checked against the same database, comparing the figures here with what a case-sensitive query would have returned. The result: the complaint set published here (1,762) is LARGER than the absolute ceiling a case-sensitive search of the same terms could have returned (876 narratives). A case-sensitive query cannot produce a bigger set than itself, so the bug is ruled out for this folder by arithmetic, not by assurance.