Glossary and Terminology

• CBT: computer-based testing, in which each exam question is not a static question but rather a random variant generated on demand from a “question template” by a Question Generator (QG)
• MCQ: multiple-choice question, whether single-answer or select-all-that-apply
• SRQ: selected-response question: any question type in which the student chooses answer(s) from among given information, rather than constructing the answer. Examples include: MCQ, true/false, matching questions.
• CRQ: constructed-response question: the student must create the response. Examples include: fill in blank, open-ended essay, numerical response.
• PL: PrairieLearn, the UIUC-developed system we’re adapting for computer-based exam generation
• QG: Question Generator—a piece of code that takes as input some specific material for a question template, generates a variant of that question based on the template, and provides the ability to check the student’s answer for that variant
• PL Element: a component with which a student interacts when answering a generated question. Examples: multiple-choice question; numeric-response question; combinational-logic circuit; “sketch a graph” question. An element includes both client-side (HTML+JS) code that handles student interaction and allows students to submit responses, and server-side code (usually Python) that interprets the student response. Often, the client-side code is expressed using XML-based markup that specifies the values of particular attributes, for example, imagine an XML description of a circuit diagram which is converted by JavaScript into the actual picture. This way, if an instructor uses only existing Elements in their QGs, the only thing the QG needs to do is set the values of those attributes in markup.