The source decides: how every answer is held to the FAA text

In a Part 107 question bank, the FAA source decides every answer, not the person who wrote the question. This is what holding a bank to that standard looks like in practice: three questions where the bank and the source were compared, what the source says, and how each question was brought into line with it and dated. It is the standard to hold any training tool to, so we show it on our own bank.

A Class E authorization question, 13 September 2026

A sectional excerpt shows a dashed magenta “keyhole” extending from a Class D circle. Your launch point is inside the magenta extension and outside the Class D circle. Is prior ATC authorisation required?

Our keyed answer was required, and the explanation cited AIM 3-2-6, which says Class E extensions to a Class D surface area begin at the surface and become part of it. That reasoning is not absurd. It is also not what governs a Part 107 authorisation.

What governs

FAA Order JO 7210.3, paragraph 19-6-3 — the order that tells the FAA’s own facilities how to process these requests — states that with regard to Class E airspace, only airspace within the lateral boundaries of the surface area designated for an airport, Class E2, requires a Part 107 authorisation or waiver. An extension to a Class C or D surface area is not that area. The FAA has separately directed that authorisations are not required for them.

So the answer is not required. Our answer was the over-cautious one, which means nobody would have flown unauthorised on our say-so. It would have been marked wrong on the exam, which is what people pay us for.

How it got in, and how it was caught

The item was written in a batch of chart questions, machine-checked against its cited section — the AIM paragraph it cites says what the explanation says it says — and had not yet been read by a person. It was found when a research document we were evaluating asserted the opposite. The document was unreliable in several other places and we nearly dismissed this claim with the rest. Instead we fetched the order and read the paragraph.

Three things followed. The question was rewritten so that the correct conclusion is available for the correct reason, and so that one distractor reaches the correct conclusion by a wrong route — because that is the discrimination worth testing. The other forty-two items in the same batch were read against their cited sections; they held. And the decoders page gained a row distinguishing a Class E surface area from a Class E extension, because if the mistake was easy for us to make it is easy for a student to make.

October 2026: the whole bank against its sources

In October 2026 every question in the bank was reviewed against the text of the FAA document it cites, and each verdict quotes the passage it rests on. The fixes shipped on 4 October 2026 as change sets 2026.10.5, 2026.10.6 and 2026.10.7. Across the three, 42 existing questions changed: 28 in the citation string only, 6 in the explanation, 2 in the question itself, and 6 in the wording of the keyed answer. One question was retired, because no FAA source in our library defines the term it tested. Two examples, in the same form as the first.

A section that does not exist. A question on dropping objects explained that "107.24 allows dropping or carriage of an object provided reasonable precautions are taken". Part 107 has no section 107.24. The rule is 14 CFR 107.23(b): no person may "allow an object to be dropped from a small unmanned aircraft in a manner that creates an undue hazard to persons or property." The sentence was in the bank from its first version. An earlier review corrected the citation string to 107.23(b) and left the explanation as it was. It was corrected on 3 October 2026, and the words "reasonable precautions" are gone, because the section does not use them.

A threshold off by one word. A weather question explained that a Convective SIGMET covers "surface winds ≥ 50 kt". The Pilot's Handbook says Convective SIGMETs "are issued for severe thunderstorms with surface winds greater than 50 knots" (FAA-H-8083-25C, p. 13-12), and the Remote Pilot Study Guide says the same (FAA-G-8082-22, p. 19). The question cited only AIM 7-1-6, which our source library does not hold, so no check had compared it with the source text. The review compared it with both documents. The explanation now says "greater than 50 knots", the keyed answer is unchanged, and the citation now names the handbook page.

The point

Every bank has errors. The measures of a tool are whether it cites sources so an error can be found, whether it corrects and dates the fix, and whether it tells you. A vendor who has never found an error in their own bank has never looked for one.

Ours is now cited to the order. The date is on this page.