Methodology

How Military Entrance Exams makes questions, checks them, and turns your answers into score estimates — so you can judge for yourself how much to trust each number.

Why there are no real test questions here

Real AFOQT, SIFT and ASTB-E questions are controlled test material. Using them would be test compromise — and the Air Force's own manual bars anyone who has taken the AFOQT from helping write a commercial AFOQT study guide (DAFMAN 36-2664, 4.12.6.1). Our questions are written without access to actual test content, from the published descriptions of each subtest. Official sample material (the AFOQT Information Pamphlet, OATTS, NMOTC samples) is used only to understand the format; it is never reproduced, and our validator rejects any question that resembles a published sample.

How a bank question is made

1. Written from scratch for one subtest and one named technique, with an explanation and a reason for every wrong option.
2. Checked automatically: five distinct options, no "all of the above", balanced answer positions, the key not habitually the longest option, no near-duplicates, and — for every math question — the key recomputed from an arithmetic expression.
3. Solved blind by a separate reviewer who sees the question without the key; only questions where the independent answer matches the key, and that the reviewer passes for clarity and accuracy, go live.
4. Logged: every question carries its provenance (author, batch, date, reviewer verdict, blind-solve match, automatic checks).
5. Watched: the share of people who get each question right, and whether strong candidates miss it more than weak ones, is computed nightly; odd questions are flagged and pulled. Every question has a "Report an issue" link.

How the visual questions are generated

Table Reading, Instrument Comprehension and Block Counting (and the SIFT's Simple Drawings, Hidden Figures and Spatial Apperception) are produced by code. Each question comes from a seed, so it can be re-created exactly to grade an answer; the answer key is computed, not typed:

• Table Reading — the key is the cell at (X, Y); every wrong option is a real misread (neighbouring cell, swapped axes, sign slip).
• Instrument Comprehension — airplanes are rendered in 3-D from a pose (heading, pitch, bank) seen from the south at the same altitude; the dials follow the reading rules in the AFOQT pamphlet; each wrong airplane differs from the correct pose in exactly one reading.
• Block Counting — piles are built from identical blocks; contacts are computed face by face (edges and corners never count); visibility is computed exactly so that every block in a pile can be seen.

Before release, automated property tests generate at least 10,000 questions of each type and check every key against an independent calculation — for example Block Counting contacts are recounted with interval arithmetic, and the drawn artificial horizon is read back to confirm it shows the intended climb, dive and bank.

How score estimates work

The services do not publish norms (AFOQT), scoring formulas (SIFT, ASTB-E) or the PCSM algorithm, so every score number in the app is an estimate from our own model, and is labelled that way. The model:

1. Your share correct on each subtest (lightly shrunk toward typical when you've answered only a few) is compared with an assumed typical performance for that subtest — the table below.
2. A composite averages its subtests (using the official composite tables) and adjusts for the fact that subtests are correlated (we assume ρ = 0.5).
3. The AFOQT result is shown as a 1–99 percentile; SIFT and OAR as 50 ± 10 on a 20–80 scale.

These assumptions are ours. Use estimates to compare yourself with yourself over time. If you share your official scores after the test (optional, in the app), we use them only to recalibrate this table.

PCSM: the illustrative estimator weights the AFOQT Pilot composite, a TBAS self-rating and flying hours (capped at 41 per DAFMAN 36-2664; an older AFPC page says 60) and shows a range, never a single "your PCSM".

Sources

• AFOQT Information Pamphlet (Form T, updated 1 Aug 2015) — subtests, timing, composites, reading rules.
• DAFMAN 36-2664, Personnel Assessment Program (17 Jan 2025, incl. Change 1, 15 Jul 2025) — minimums, retakes, superscoring, E-AFOQT, PCSM, study rules.
• AFPC PCSM program pages and FAQs — PCSM inputs, TBAS, flying hours, superscoring.
• Frequently Asked Questions about the SIFT (V2.0) and Fort Novosel's Army Aviation FAQs — SIFT subtests, scale, retakes.
• Navy Medicine Operational Training Command ASTB-E Overview and FAQ — ASTB-E subtests, scores, retakes.

Found something out of date? Tell us — facts are re-checked whenever a source changes.