Guide-linked quiz · Medium Level

AI for Systematic Review Searching Quiz

Test AI systematic-review workflows: search reproducibility, active-learning screening, missed records, human review, extraction validation, and PRISMA

Question 1 of 8

What does an active-learning screening system typically do after a reviewer labels records?

Preview every question8 questions

Answers and explanations appear as you take the quiz.

  1. What does an active-learning screening system typically do after a reviewer labels records?
    • Automatically declares all unshown records ineligible
    • Uses those labels to prioritize the next records likely to be relevant
    • Replaces the review protocol
    • Generates the final evidence synthesis without source checking
  2. Why should teams check low-ranked records when using AI prioritization?
    • Treat every low-ranked record as excluded once a set count is reached
    • Stop when a number of consecutive irrelevant records is reached without checking recall
    • Relevant studies may still be ranked low and could be missed by an early stop
    • Assume the classifier’s confidence proves an article is ineligible
  3. Which automation detail does PRISMA 2020 ask reviewers to report when applicable?
    • Only the model vendor’s name
    • Only the final number of included papers
    • The number of pages in each article
    • How many reviewers screened records, whether they worked independently, and automation tools used
  4. A language model extracts an outcome value from a paper. What should the reviewer do?
    • Verify the value against the source table or text and record its location
    • Ask the same model to verify its own extracted value
    • Compare only with an abstract-level summary, not the paper
    • Infer the value from the article title and study population
  5. Why preserve exact database search strings and dates?
    • So an AI model can change eligibility criteria
    • So another team can reproduce or update the search
    • To prove that all relevant studies were found
    • To replace the protocol
  6. What does a stopping rule based on consecutive irrelevant records require?
    • A guarantee that recall is 100 percent
    • A replacement for full-text screening
    • Assessing and reporting the risk that eligible records remain below the threshold
    • No human validation
  7. How should AI-generated search synonyms be treated?
    • As automatically valid for every database
    • As a reason to skip controlled vocabulary
    • As evidence that the search is comprehensive
    • As candidates to test with domain expertise and document in the final strategy
  8. Who remains responsible for the review’s coverage and conclusions?
    • The review authors
    • The screening software vendor
    • The database provider
    • The model that ranks abstracts

Keep testing yourself

More quizzes picked for your level.

View all quizzes