Formative assessment is an assessment placed tactically in the learning activities of a course to help the learning process. If a numerical mark is produced, it does not contribute to grading of the student. Some support may be provided via automated feedback and some may be provided by teachers after analysing performance.
Summative assessment is an assessment that is used to determine the success of the student in attaining the published learning objectives of the course and will contribute to grading.
Practice summative assessment is an assessment that doesn't contribute to grading of the course but models the real summative assessment so that students can measure their own ability using the same instruments as those used for formal assessment.
Computer delivered assessment is an assessment that requires the student to use a computer to enter responses.
Computer marked assessment is an assessment that students may complete using pens or pencils with special cards or sheets which are later marked using specialist hardware and computer software.
An invigilated examination is conducted in a room which may or may not be equiped with computer and has trusted staff who guarantee that students do not copy each other's work or use aids which are not permitted.
Plagiarism in the context of multiple choice refers to copying another student's answers or the use of prohibited reference material to answer questions.
Plagiarism detection services rely on pattern matching of plain text content against vast databases of text. They can therefore only be applied to essays, theses, short written answers, project reports etc. There is no difference between the responses of the honest a able students in the class on a multiple choice exam. A dishonest student who cheats effectively cannot be distinguished from this group. This advice applies to all delivery mechanisms and has nothing to do with the use of computer delivered assessment versus paper delivered assessment and has nothing to do with the individual properties of the various software products - it is a result of the fundamental design of the question.
Students who score badly on an MCQ examination are expected to give different answers. If a set of students give the same wrong answers they may have cheated. However, the evidence is much weaker than you would have for plagiarism on essays and it may be difficult to separate victim from offender.
This follows from the above points and applies to computer delivered assessments using any software products and paper delivered assessments. Because of the possible stresses induced by the basic unreliability of computer hardware it may be best to use computer marked assessments that are delivered through paper/card. However, software for marking paper exams often has less question types and marking schemes than are available in computer delivered assessments.
Work in progress......