The GMAT is a computer-adaptive standardized test of reasoning that is used by B-schools to compare applicants. The key word here is adaptive, as the level of difficulty of each subsequent question depends on how the test taker has performed on the previous question/s. While both the Quantitative and Verbal sections are adaptive, the Integrated Reasoning Section is NOT computer adaptive.
At the start of the Quant / Verbal section of the exam, the test taker is presented with a question of medium difficulty. As the test taker answers each question, the computer scores the chosen answer and uses it - as well as the other responses given to any preceding question - to determine which question type to present next and the difficulty level of the next question. Because the computer scores each question before selecting the next one, you cannot skip; return to; or change your responses to previous questions.
Correct responses typically prompt questions of ‘increased’ difficulty, and incorrect responses generally result in questions of ‘decreased’ difficulty. This process will continue until the test taker completes the section, at which point the computer will have an accurate assessment of the ability level of the test taker.