Guide-linked quiz · Medium Level

Predictive Policing and Feedback Loops Quiz

Explain how police attention can shape recorded data, distinguish place- and person-based tools, and evaluate claims about predictive policing.

Question 1 of 8

What distinguishes place-based from person-based predictive policing?

Preview every question8 questions

Answers and explanations appear as you take the quiz.

  1. What distinguishes place-based from person-based predictive policing?
    • One uses cameras and the other uses radios
    • One forecasts locations; the other ranks individuals
    • One is always legal and the other is banned
    • There is no difference
  2. Why can arrest data be a problematic proxy for underlying crime?
    • Arrests reflect both behavior and enforcement patterns
    • Arrests are never recorded
    • Models cannot use counts
    • Arrests always understate crime equally
  3. How can a feedback loop arise?
    • A model is deleted after each shift
    • Police records are always independent of deployment
    • A map is printed twice
    • Patrol forecasts change where officers look, and resulting records feed later forecasts
  4. What status does Chicago’s public historical dataset give the Strategic Subject List program?
    • It is mandatory nationwide
    • It is current and in active deployment
    • The program ended in 2019 and the dataset is retained for historical reference
    • It never used individual scores
  5. What did the NIJ-sponsored 2016 evaluation find about victimization among SSL-listed people?
    • They were all victims of shootings
    • The list reduced city homicide by half
    • The study did not include a comparison group
    • They were not more or less likely than a comparison group to become homicide or shooting victims
  6. Why should analysts account for concurrent policy changes when evaluating a tool?
    • They can also change investigations and outcomes
    • They guarantee the model is unbiased
    • They remove the need for data
    • They change the law of probability
  7. Does a model score prove that a person committed or will commit a crime?
    • Yes, if the score is high
    • No; it is a forecast, not a factual finding
    • Only if police record it
    • Yes, in person-based tools
  8. What helps evaluate claimed crime reduction?
    • Only compare predictions with themselves
    • Count the number of map tiles
    • Measure outcomes independent of police activity and account for alternative explanations
    • Use model accuracy alone

Keep testing yourself

More quizzes picked for your level.

View all quizzes