PULSCAR ANALYSIS OF NHTSA PUBLIC DATA — WHEEL ALIGNMENT 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. 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 campaign number. Field definitions: https://static.nhtsa.gov/odi/ffdd/rcl/RCL.txt Communications FLAT_TSBS.txt manufacturer communications; counted DISTINCT on the NHTSA document id. Field definitions: https://static.nhtsa.gov/odi/ffdd/tsbs/TSBS.txt A complaint is an owner's voluntary account, not a verified repair and not a failure rate, and there is no denominator of vehicles on the road. Nothing here ranks reliability. One selection effect matters especially for this topic. People file with a federal safety agency when something worries or costs them. A correct, uneventful $130 alignment never appears in this file. What does appear is alignments that were repeated, alignments that did not work, and cars that would not stay aligned. That makes the file useless as a measure of how often alignments succeed — and unusually good at showing what happens when alignment is sold as the answer to something else. -------------------------------------------------------------------------- 2. THE COMPLAINT QUERY (4,951 -> 3,590) -------------------------------------------------------------------------- Step 1 - FTS5 search for alignment OR aligned OR realign OR realigned OR realignment OR aligning -> 4,951 complaints. This catches every sense of the word, including misaligned doors, tailgates and headlamps. Step 2 - wheel context. At least one occurrence of align* had to sit within 120 characters of one of: wheel | tire | tyre | front end | front-end | suspension | steering | camber | caster | toe | thrust -> 3,590 complaints. Sample reading: 28 random complaints from the final set read in full. 25 were about wheel alignment in some capacity. 3 were not — one listed "wheels were not aligned" among a dozen unrelated used-car defects, one attributed alignment trouble to a power steering failure without a wheel alignment ever being performed, and one used the word only to say the problem was NOT alignment. Measured false-positive rate: roughly 11% (3/28). Every count here carries it. Note that the query deliberately keeps complaints where the word appears because a shop recommended, performed or blamed an alignment. That is the subject of this research, not an error. -------------------------------------------------------------------------- 3. MILEAGE -------------------------------------------------------------------------- Owner-entered failure mileage, with values <= 0 or >= 400,000 dropped. 1,924 of 3,590 complaints (53.6%) carry a usable mileage, so every median describes that subset and nothing corrects for the rest. Whole set: median 52,000 miles; 18.9% at or under 12,000 miles; 38.9% at or under 36,000. -------------------------------------------------------------------------- 4. CONTEXT SUBSETS (alignment-complaints-by-context.csv) -------------------------------------------------------------------------- Regexes over the narrative. Subsets overlap and do not sum to the total. pulls/drifts pull* to|toward|hard|right|left | wander* | drift* to|toward|right|left | veer* new tires new tire | new set of tire | replac* (the) tires | four new tire | bought tires uneven wear uneven (tire|tyre|tread) | tire(s) wore/worn/wear unevenly|out|prematurely|fast | inner/outer/inside/ outside edge | feather* | cupp* | premature* tire | scallop* after suspension align* within 120 chars of tie rod | control arm | or steering work ball joint | strut | shock | knuckle | subframe | steering rack | bushing (either order) spec/adjustability out of|outside (factory|manufacturer's) spec* | can't|could not|unable to be aligned | no (factory) adjust* | not adjustable | cannot|could not be adjusted | beyond spec | no way to adjust ADAS lane keep | lane depart | lane assist | adas | calibrat* | steering angle sensor | forward collision | adaptive cruise | camera pothole/curb pothole | curb | kerb off-center off-cent* | crooked | steering wheel is/sits/was turned|tilted|angled recall mentioned recall* | campaign number Two subsets were built with more care because the article leans on them. REPEATED (189 complaints). The narrative had to say the alignment happened more than once: align* followed within three words by twice | three times | four times | multiple times | several times | repeatedly | again; or (second|third|another|repeated|multiple|several|numerous) [wheel | front end | 4 wheel] align*; or re-align* the (vehicle|car|truck|wheels); or aligned (it) twice|again|multiple times|several times|repeatedly|N times. An earlier version of this pattern allowed a bare digit near "align" and matched the phrase "4 wheel alignment" on almost every complaint in the file. That version was discarded; the digits were removed from the pattern. INEFFECTIVE (116 complaints). align* within 140 characters of a negated fix verb (did not / didn't / doesn't / wouldn't / couldn't / failed to / not + fix | solve | correct | help | resolve | cure | stop | alleviate | eliminate); or still|continued to|continues to|kept + pull*|drift*|wander*| wear* within 140 characters of align*; or align* followed by "but it/the still", "problem persisted|remained|continued|returned". UNION (292 complaints) is the distinct union of the two, and is the set used for alignment-what-it-turned-out-to-be.csv. Sample reading of those two subsets: 12 random complaints read from each. All 24 were genuinely about an alignment that was repeated or that failed to fix the reported symptom. No false positives were found in that sample, which is a small sample and is reported as one. Baseline row: the towed and crash shares of all 777,040 complaints in the file (5.11% towed, 4.37% crash). -------------------------------------------------------------------------- 5. WHAT IT TURNED OUT TO BE (alignment-what-it-turned-out-to-be.csv) -------------------------------------------------------------------------- For each of the 292 repeated-or-ineffective complaints, we counted whether the narrative names another component. Regexes: tires tire(s) within 60 chars of defect|conicity|bad|replac*| swap*|new|brand|warrant*, or "radial pull" suspension strut | shock absorber | shocks | control arm | ball joint | bushing | tie rod | track bar | steering damper | sway bar balance/rim balanc* | bent rim | bent wheel | road force steering steering rack|gear|column|coupler|shaft | rack and pinion | electric power steering | eps | power steering death wobble death wobble structure subframe | sub-frame | cradle | frame | chassis | unibody unidentified could not be duplicated|determined|identified|found | no cause/problem was found | unable to duplicate|determine| identify | never found|fixed|identified IMPORTANT LIMIT: these count MENTIONS, not confirmed root causes. A complaint that says "they replaced the struts and it still pulled" counts in the suspension row without the struts having been the cause. Categories overlap and do not sum to 292. The table shows what else was in the room, not a diagnosis. -------------------------------------------------------------------------- 6. MAKE AND MODEL (alignment-complaints-by-make-and-model.csv) -------------------------------------------------------------------------- Rows with fewer than 15 complaints are omitted. "level" is make or model. Report counts, not failure rates, not a ranking. -------------------------------------------------------------------------- 7. RECALLS (alignment-related-recalls.csv) -------------------------------------------------------------------------- Query: recall rows whose defect text contains "wheel alignment", "alignment adjust", "camber", "toe angle" or "thrust angle", or whose consequence or remedy text contains "wheel alignment". Counted DISTINCT on campaign: 20 campaigns. A broader search on the bare word "alignment" anywhere in the defect, consequence or remedy returns 74 campaigns, but most of those are misaligned tailgates, door handles and headlamps rather than wheel geometry. The narrower query is the one published; the broader number is noted here so nobody is surprised by the difference. The file includes equipment recalls and non-passenger vehicles (a bike carrier whose product name is "Camber", a school bus camber-caster kit bolt). They were left in rather than silently filtered and are visible in the makes column. -------------------------------------------------------------------------- 8. MANUFACTURER COMMUNICATIONS -------------------------------------------------------------------------- Query: FTS5 on the bulletin summary for alignment OR camber OR caster OR toe, counted DISTINCT on the NHTSA document id: 808 documents. Narrower counts used in the article: 15 documents whose summary names alignment and tire wear; 22 whose summary names both alignment and calibration. No CSV is published for these; the bulletins cited in the article are linked individually to their own NHTSA PDF. -------------------------------------------------------------------------- 9. LIMITATIONS, IN ONE PLACE -------------------------------------------------------------------------- - Complaints are self-reported and unverified. - There is no denominator. Nothing here is a failure rate, and nothing here measures how often alignments work. - The file is selected toward alignments that went wrong or were sold wrongly; successful ones are invisible to it. - Keyword sets carry the false-positive rate measured in section 2. - 46.4% of the set has no mileage; medians describe the rest. - Subsets overlap and do not sum. - Section 5 counts mentions, not causes. - Recall "potentially affected" counts are population estimates. Questions or a correction: support@pulscar.io ------------------------------------------------------------------------------ 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 (4,951) is LARGER than the absolute ceiling a case-sensitive search of the same terms could have returned (2,427 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.