To raise your AFQT, study the 4 subtests it is built from: Arithmetic Reasoning, Mathematics Knowledge, Word Knowledge and Paragraph Comprehension. Give Word Knowledge a slot every day, because each point you gain there adds about 1.3 points to the AFQT composite, more than any other subtest. Then practice at the real time limits until you can finish every subtest, since the CAT-ASVAB penalizes unfinished items.
Section 2.4 of the 1997 score scale report by Daniel Segall gives the AFQT as a plain sum of rounded standard scores: AR + MK + 2 x VE, where AR is Arithmetic Reasoning, MK is Mathematics Knowledge and VE is Verbal Expression, a blend of Word Knowledge (WK) and Paragraph Comprehension (PC). Each standard score sits on a scale with a mean of 50 and a standard deviation of 10. The composite runs from 109 to 269 in the conversion table, and Table 2.5 turns it into the percentile you see on your score sheet: the percentage of the 1997 national youth sample who scored at or below you.
VE is the part most people get wrong. It is not WK + PC. On the current scale, VE is an optimally weighted composite in which Paragraph Comprehension receives 0.61934 times the weight of Word Knowledge, restandardized to mean 50 and rounded to a whole number. In standard-score terms, VE comes out to roughly 0.655 x WK + 0.406 x PC, minus about 3. Double those weights and you have what 1 point on each subtest is worth in the composite. The Verbal Expression guide shows how a strong WK with a weak PC, or the reverse, plays out.
"VE counts double" is true but easy to misread. The double weight is shared between 2 subtests, so the verbal half and the math half of the AFQT are close to balanced. Work on the half where your scores have room to move; verbal does not always win.
Take a profile of 45 on all 4 AFQT subtests. VE is 45, the composite is 45 + 45 + 90 = 180, and Table 2.5 converts 180 to the 28th percentile. The table below adds 5 standard-score points to 1 or 2 subtests at a time and runs each profile back through the same formula and conversion.
Gaining 5 points on both verbal subtests lands exactly where gaining 5 on both math subtests does: a composite of 190 and the 38th percentile. On its own, 5 points of Word Knowledge reaches the 34th percentile, ahead of Arithmetic Reasoning at 33 and Paragraph Comprehension at 32.
Rounding also matters. VE is rounded before it is doubled, so a 1-point change on one verbal subtest sometimes moves nothing.
The percentile table is not a straight line either. Near the middle of the distribution it is steep: a composite of 200 is the 48th percentile and 210 is the 59th. Near the top it flattens, with 240 at the 87th and 250 at the 93rd. The same 10 composite points buy about 11 percentile points in the middle and about 6 at the top.
A 5-point gain is half a standard deviation on that subtest, a large change. The table shows the arithmetic, not a forecast.
The move from 28 to 38 crosses a line that matters. The Army listed 31 as its enlistment minimum, and the Marine Corps and Air Force listed 31 for high school graduates, on their recruiting pages as of October 2026. Minimums change, so confirm yours with a recruiter.
If you have already tested, your score sheet lists a standard score for each subtest. If not, take a timed practice set on the 4 AFQT subtests and note where your misses cluster. Weight your plan toward your lowest subtest, and keep a daily Word Knowledge slot either way, since it pays the most per point.
Match the session to the subtest. Word Knowledge is 15 scored items in 9 minutes, and vocabulary grows through short daily repetition rather than long sessions. Arithmetic Reasoning gives 55 minutes for 15 word problems, each of which you set up before solving, so it needs longer blocks. Paragraph Comprehension gives 27 minutes for 10 items and rewards careful reading over speed. The Word Knowledge guide covers how those items are built.
The plan assumes 60 to 70 minutes a day, 6 days a week, with the 7th day for rest or catching up. Day 1 is your baseline: a timed practice set on all 4 AFQT subtests. Every block ends the same way: go back over each miss and write down why the right answer is right. Untimed practice builds skill, but from week 3 on, run at least one block a day under the real time limit.
In week 4, take one full timed practice test early in the week and spend the following days on what it exposed. Keep the last 2 days to light review. If your baseline shows math far behind verbal, cut vocabulary to 15 minutes and give math 40. With more than 4 weeks, the longer ASVAB study plan spreads the same content out.
The computer adaptive version, CAT-ASVAB, is the one given at Military Entrance Processing Stations and most Military Entrance Test sites. Technical Bulletin No. 3 describes the cycle: every examinee starts at an ability estimate of 0, which is medium difficulty, and gets the item that is most informative at that level. After each answer the system updates its estimate and picks the next item. A correct answer leads to a harder item, a wrong one to an easier item. The guide to how the CAT-ASVAB works covers the item response theory behind each pick.
Early in a subtest the estimate rests on only a handful of answers, and every next item is chosen from that estimate. A careless miss on item 2 sends you to an easier question, and you spend the next few items working back up. The program publishes no rule that early items count more; your final score comes from the difficulty and correctness of all your answers. What early accuracy buys is a path through harder, more informative items. You also cannot go back: once you answer an item and move on, the answer is final.
If time runs out, section 4.1 of the bulletin applies a penalty sized to the number of unfinished items, equivalent to the score you would get by guessing at random on them. The program's preparation advice adds that the CAT-ASVAB applies a relatively large penalty when several incorrect answers come at the end of a subtest, and tells you to keep answering properly rather than filling in random guesses. Neither blanks nor random clicks help.
Most people have enough time. The program says nearly all examinees finish each subtest, and in the 2008 equating study 99.0 percent finished Arithmetic Reasoning on Form 9R under a 39-minute limit, shorter than today's 55 minutes. The ASVAB test day guide shows where the clock and the remaining-question count appear on screen, the 2 numbers you pace against. The paper version works differently: you can review answers within a subtest, and guessing carries no penalty.
The official retest policy sets the waits below, and they apply to school and enlistment tests alike. A calendar month runs date to date: the policy's own example is a first retest on March 20, which allows a second retest on or after April 20.
The Army states that your most recent score, not your highest, is the one used, so a rushed retest can replace a 38 with a 33. Book the next attempt only when your timed practice scores are consistently above your current result, not when the wait runs out. Scores stay valid for enlistment for 2 years from the test date.
The first mistake is studying the wrong subtests. General Science, Electronics Information, Auto and Shop Information, Mechanical Comprehension and Assembling Objects never enter the AFQT. They feed the line scores that qualify you for specific jobs, which matter once you clear the AFQT minimum, not before.
The second is trusting guessing patterns. Tips such as "pick the longest answer" or "go with C" appear nowhere in the official preparation advice, which tells you to rule out as many wrong choices as you can and make an educated guess from the rest. On the CAT, a run of wrong guesses at the end of a subtest is exactly what the scoring penalizes.
The third is skipping Paragraph Comprehension because it carries less weight than Word Knowledge. It still adds about 0.8 composite points for every point you gain. Check every choice against the passage itself: an answer that is true in the world but not supported by the passage is still wrong.