The percentages of FY 1999 active duty NPS accessions in each AFQT category are shown in Table 2.8. The percentage of recruits in Categories I and II was approximately the same as their civilian counterparts (males - 39 versus 39 percent; females - 34 versus 33 percent). Category III accessions greatly exceeded civilian proportions (males - 61 versus 30 percent; females - 66 versus 37 percent), while the percentage of recruits in Category IV was much lower than in the civilian population (males - 1 percent versus 20 percent; females - less than 1 percent versus 22 percent). The low percentage of Category IV recruits is, in part, a result of DoD limits of 4 percent Category IV recruits, with even lower Service limits. Ten percent of civilian males and 9 percent of civilian females scored in Category V; DoD allows no Category V recruits.
The paper-and-pencil version of the test administered at a Mobile Examination Test (MET) site usually takes 3–4 hours. The time needed to take the CAT-ASVAB test can vary. The test is adaptive. If the candidates answer a question correctly, they are given one of increasing difficulty. If the candidates miss a question, they are subsequently given an easier item. This pattern continues until the test is finished. Because of its adaptive nature, the CAT-ASVAB test generally takes about half the time of the paper-and-pencil version.
After adding so much information to the aptitude test, there was a bit of difficulty interpreting the test results. In addition to that, a vast majority of test takers were deemed as being under qualified based on their test results. This is why the percentile change was made. It ensured that a 50% actually correlated with a person doing better than 50% of the test takers. Those revisions have worked hand-in-hand with the preparation of the armed forces.