A question bank is a large set of practice questions, each with a worked explanation, tagged by subject and topic, so that a candidate can practise the areas they keep getting wrong. Question banks are the main product for many professional exams, such as the licensing exams for doctors, nurses and accountants. Candidates often call them Q-banks.
A good question bank has three things. First, many questions written in the style of the real exam, often by doctors or accountants who have passed it. Second, an explanation for every option: why the right answer is right, and also why each wrong answer is wrong. Third, tags. Each question is tagged with its subject, its topic and its difficulty. The tags let the software build a set of questions on any topic, and show the candidate where they are weak.
Consider a young doctor in Kolkata preparing for NEET PG, the exam for postgraduate seats in medicine. Her question bank has 15,000 questions across 19 subjects. She practises in blocks of 40 questions, timed like the real exam. After 1,200 questions, her dashboard shows that she gets 71% right overall, but only 48% in pharmacology. Her worst topic is antibiotics, where she has got 22 of 40 questions wrong.
So she builds a custom block: 40 questions, only on antibiotics, and only ones she has not seen before. When she gets one wrong, she reads the explanation. It shows why the drug she chose is wrong for that patient. The software also marks the question to show her again a week later. Two weeks on, her pharmacology score is 63%.
Behind the screen, the software keeps every answer that every candidate gives, and uses those answers to judge each question. If 95% of candidates get a question right, it is easy. If strong candidates get a question wrong as often as weak ones, it may be badly written, and an editor checks it. The same data lets the software tell each candidate how she compares with everyone else who uses the bank.
Question banks also feed other features. Online mock tests draw their questions from a bank. AI doubt solving works from the bank's checked explanations, so that its answers rest on a reliable source. An AI tutor can now write new questions in the bank's style, but an expert must check each one before it is added. Most test prep for GRE, GMAT and licensing exams is built around a question bank.