The short version
The ASVAB is the Armed Services Vocational Aptitude Battery, the test taken by everyone entering the United States military. It is a battery rather than a single exam: nine separate subtests, each with its own subject and its own clock, producing a set of scores rather than one number.
Almost everything people believe about it comes from confusing the battery with one score inside it.
Only four subtests decide whether you can enlist
This is the fact worth knowing before anything else. The number that determines enlistment eligibility is the AFQT, the Armed Forces Qualification Test, and it is computed from four subtests only:
- Arithmetic Reasoning
- Mathematics Knowledge
- Word Knowledge
- Paragraph Comprehension
General Science, Electronics Information, Auto and Shop Information and Mechanical Comprehension contribute nothing to it. A candidate can score brilliantly on all of those and still fail to qualify, and a candidate can score zero on all of them and enlist comfortably.
The other subtests are not pointless: they feed the line scores that decide which jobs you qualify for, and for many people that is the more important question. But if the goal is simply getting in, four subtests carry the whole weight, and preparation time should be divided accordingly rather than spread evenly across nine.
The AFQT is a percentile, and people misread it constantly
An AFQT of 50 does not mean you answered half the questions correctly. It means you performed better than 50% of a nationally representative reference group of 18 to 23 year olds. The scale runs from 1 to 99.
This trips people up in both directions. A candidate who feels they did badly can score respectably, because the comparison group includes everyone, not just the motivated. And a candidate who found the questions easy can be disappointed, because so did everybody else.
Word Knowledge and Paragraph Comprehension are combined into a Verbal Expression composite before the AFQT is computed, and in the standard formula that verbal composite is counted twice while each mathematics subtest is counted once. Vocabulary carries more weight in the enlistment score than most candidates assume, and vocabulary is the single most improvable thing on the whole battery.
Our nine subtests, and what each one is
Our practice forms carry 225 questions: 35 Word Knowledge, 30 Arithmetic Reasoning, 25 Mathematics Knowledge, 15 Paragraph Comprehension, 25 General Science, 25 Auto and Shop Information, 25 Mechanical Comprehension, 25 Assembling Objects and 20 Electronics Information.
Note the shape of that list. Word Knowledge is the largest subtest on the paper, and Paragraph Comprehension the smallest, and both are AFQT subtests. That is a strong hint about where study time earns most.
The computer version does not let you go back
Most candidates sit the CAT-ASVAB, the computer-adaptive version. It chooses each question based on how you answered the last one, which means it cannot let you revisit an answer: the question you are looking at only exists because of the one before it.
Practically, that changes the strategy completely. On a paper test the standard advice is to skip the hard ones and come back. On the CAT there is no coming back. You answer the question in front of you, and you accept that answering it well matters more early in the subtest than late, because the early questions steer the difficulty of everything that follows.
The unproctored version is a trap with a good side
The PiCAT is a version of the ASVAB you can take at home, unproctored, on your own time. It sounds like the easy route and it is not, because a PiCAT result has to be confirmed by a shorter proctored verification test afterward. If your verification does not line up with your PiCAT, the PiCAT is thrown out and you sit the full battery instead.
Used honestly it is genuinely useful: it gets the long part out of the way in comfortable surroundings and leaves a short confirmation on test day. Used as a shortcut it costs you the whole thing and a fair amount of goodwill. Whether it is offered at all is up to your recruiter.
The scores follow you further than the test does
ASVAB results are used for enlistment, for job assignment, and they remain valid for a period afterward, which means a score taken casually in high school can still be the number on the table years later when someone is serious about a particular role. Line scores are built by combining subtests in different combinations for different jobs, so a strong Electronics Information result can open a specific door that nothing else opens.
The practical consequence: decide what you want to do before you decide how to prepare. "Get a good ASVAB score" is not a plan, because there is no single ASVAB score.
What we found when we tested our own exam
Free ASVAB practice tests are everywhere and hardly any of them have been measured for the oldest flaw in multiple choice: a correct answer that is reliably longer than the wrong ones. When that happens, a candidate can beat the paper without reading it, and then walk into the real battery with a number that meant nothing.
So we measure it. Across the 16 sections of our ASVAB exams long enough to measure (420 questions, in sections of 20 or more), the best score any length-based strategy reaches is 38%, against the 25% that guessing alone should give. Four strategies are scored: pick the longest option, pick the shortest, cross off the longest and guess, cross off the shortest and guess.
Adults sitting the ASVAB are the group most likely to be working from a free practice test of unknown provenance, which is exactly the situation where an unmeasured bank does the most damage.
That figure is generated from the live bank on every deploy, and the build fails if any section crosses 40%. Our own ASVAB forms were among the offenders once: three sections across the two forms sat over the line in a backlog of 46 sections and 1,741 questions, and one of them regressed after a fix made in the database was never written back to the file it loads from. The check now runs at load time too, for exactly that reason.
How to prepare, in order
Start with vocabulary, because Word Knowledge is the largest subtest, it feeds the AFQT, it is double-weighted through the verbal composite, and it is the one part of this battery where a few weeks of flashcards genuinely moves the number.
Then Arithmetic Reasoning, which is word problems rather than computation and where most losses come from setting the problem up wrongly, not from the arithmetic.
Then Mathematics Knowledge, which is straightforward school algebra and geometry and responds to ordinary revision.
Only then the technical subtests, and only the ones your target job needs. Studying Auto and Shop Information for a role that does not use it is time taken from the four subtests that decide whether you get in at all.
Our shorter ASVAB guides go deeper on the four AFQT subtests, how the battery is scored and the CAT against the paper version.
Where to practice
Our two ASVAB forms carry all nine subtests and 225 questions, and you can sit a single subtest on its own when only the four AFQT ones matter to you. Every wrong answer carries its own explanation saying what a student who picked it was probably thinking, which is the part that turns a score into something you can act on. There is a free sample on the samples page if you want to see the format before paying for anything.
Common questions
What is the AFQT and how is it different from the ASVAB?
The ASVAB is the whole battery. The AFQT is a single score computed from just four of its subtests: Arithmetic Reasoning, Mathematics Knowledge, Word Knowledge and Paragraph Comprehension. The AFQT decides whether you can enlist; the rest of the battery decides what you can do once you have.
Is the AFQT a percentage?
No. It is a percentile from 1 to 99, comparing you against a reference group of 18 to 23 year olds. An AFQT of 50 means you scored better than half of that group, not that you got half the questions right.
What is a good ASVAB score?
It depends entirely on what you want to do. Each branch sets its own minimum AFQT for enlistment, and each job sets its own line-score requirements from the other subtests. A score that easily enlists you may not open the job you want.
Can you go back and change an answer on the ASVAB?
Not on the computer-adaptive version, which is what most people sit. Once you submit an answer it is final, and the next question is chosen based on it. The paper version does allow review within a subtest.
How many questions are on the ASVAB?
It varies by version. Our practice forms carry 225 questions across nine subtests, which mirrors the paper battery.
Practice the real thing
Our ASVAB exams are full length and written from scratch — every wrong answer carries its own explanation, so a child learns what they got wrong and why. Ten questions are free, with no account and no card. After that it is five credits an exam.