False logs vs. ELD tampering: the same rule, 29% vs. 97% out of service

49 CFR §395.8(e) is one paragraph with two halves. Paragraph (e)(1) says no driver or carrier may make a false report of the driver's record of duty status. Paragraphs (e)(2) and (e)(3), the ones FMCSA started coding on April 1, 2026, say nobody may tamper with an ELD so that it does not accurately record and retain the required data. Both are "falsification" in everyday language. In the inspection data they are different codes, written at very different rates, with very different outcomes. Our ELD Tampering report covers the second half only; this article puts the two side by side.

Method: every Part 395 violation row whose section starts with 8E or 8(e), joined to its inspection. Rows coded 8E-HOSPDELDT or 8E-HOSPVELDT are the tampering group; every other 8E section is the false-log group. An inspection counts once per group, and "out of service" means at least one row of that group in that inspection carried the OOS flag. The window is the six months before the tampering codes existed plus the time since: October 1, 2025 through October 12, 2026, the last date in the feed.

Twelve codes, two families

"OOS rows" counts violation rows with the out-of-service flag, not inspections; one inspection can carry several rows of the same code. Descriptions are FMCSA's, trimmed. Source: FMCSA roadside inspection and violation data (open data via Socrata) in the Carrierintel database, Part 395, inspections 2025-10-01 → 2026-10-12, computed 2026-10-14.

The false-log family is ten codes and it splits by what the false record did. 8E-HOSPD — a false entry that did not conceal an hours violation — is the bulk: 35,410 inspections in the window, and only 3,961 of its 62,517 rows are out of service. 8E-HOSPDOOS — a false record that concealed a limit violation since the last rest — is the other way round: 9,573 of 13,676 rows are OOS. The personal-conveyance codes (8E1PC, 8E1PC-OOS, 8E1-PC) are a false log of a specific kind, driving off-duty that was not; since April they are 5,310 of the 30,526 false-log inspections, 17.4%. The tampering family is two codes, property and passenger, and 12,972 of its 14,107 property rows are out of service.

Since April: four times as many false logs, less than a third of the out-of-service rate

Since April 1, when both codes have existed, inspectors wrote a false-log citation on 30,526 inspections and a tampering citation on 7,149 — a ratio of 4.3 to 1. The outcomes are the mirror image: 8,865 of the false-log inspections (29.0%) ended with the driver out of service, 6,926 of the tampering inspections (96.9%) did. 21,080 carriers were cited for a false log, 5,040 for tampering. Against every inspection that had any Part 395 violation in the period — 131,871 — the false-log family is 23.1% and tampering 5.4%.

They rarely appear together. Of the 36,818 inspections that cited either, 857 cited both, 6,292 cited tampering only and 29,669 cited a false log only. So 12.0% of tampering inspections also carried a false-log row, and 2.8% of false-log inspections also carried a tampering row. An inspector who writes tampering usually writes it alone.

The false-log count did not fall when the tampering codes arrived

Tampering columns are empty before April 2026 because the codes did not exist. October 2026 covers the 1st to the 12th only.

One question the April codes raise is whether the false-log count dropped when a tampering code became available. In the totals it did not: false-log inspections ran at 4,368 to 5,780 a month in the six months before April (29,015 in total) and at 4,545 to 5,108 a month in the six months after (29,221), and the out-of-service share stayed between 27.1% and 31.1% the whole time. The tampering series started at 869 in April, stepped up to 1,281 in June and has stayed between 1,171 and 1,308 a month since. Monthly totals cannot show whether individual citations moved from one code to the other; they show that the false-log series kept its level while the tampering series appeared beside it.

By state: who writes false logs, who writes tampering

Federal (non-state) inspections are excluded. States are ranked by false-log inspections; a dash means the state wrote no tampering inspections, so there is no share to show.

The two maps are different. California and Texas write the most false-log citations (2,793 and 1,835) and few tampering ones (248 and 103). Among these twenty states, Arizona, Alabama and Oregon are the only ones where tampering inspections outnumber false-log inspections — 1,899 to 1,812, 908 to 858, 1,225 to 745. Illinois wrote 810 false-log inspections, with the highest out-of-service share in the table at 58.4%, and zero tampering inspections. Tennessee's false-log OOS share is 47.3%, Indiana's 13.1%: the same code family, written very differently from one state to the next. The full 43-state tampering table, with its out-of-state shares and monthly series, is in ELD tampering by state.

What this means when you read a carrier's record

Methodology

Source: FMCSA roadside inspection and violation data via Socrata, imported into the Carrierintel database — the raw violations and inspections tables, not the ELD tampering cache. Part 395 only; sections starting 8E or 8(e); 8E-HOSPDELDT and 8E-HOSPVELDT form the tampering group (no 8(e)(2) or 8(e)(3) section occurs in the feed), every other section the false-log group. Window: inspection dates 2025-10-01 through 2026-10-12, the last date in the feed on 2026-10-14; the "since April" figures start 2026-04-01. The raw feed runs one day past the report cache, so the tampering totals here (7,149 inspections, 5,040 carriers) sit slightly above the ELD Tampering report figures of the same day (7,117 and 5,024). Computed 2026-10-14 on the production database; SQL published beside the article as false-logs-vs-eld-tampering-2026.sql.