• In cases where X cannot be manipulated = random assignment is impossible or ethically unacceptable (e.g., political party affiliation, personality, socio-economic status, disease etc.)
  • Natural occurring groups are measured and taken as experimental conditions
    • ! Causal interpretations should be made with caution !
  • The effect could be due to a confounding variable, which just coincides with quasi-X
  • Disadvantages of Laboratory (Quasi-)Experiments
  • External validity is not guaranteed

Politics

  • this is especially hard with politics
  • e.g. right-wing people do not support issue X
    • might depend on other closely correlated traits
      • educational level
      • income bracket
      • occupation
      • upbringing
      • etc…

Quasi Field Experiments

  • A controlled experiment which occurs at places with people going about their normal behavior (e.g. school, parking ramp)

  • When trying out behavioral interventions in real-world contexts Field Experiments are often Quasi- Experiments

    • A team class gets the intervention, the other team gets no intervention, or an other intervention with a known effect, or gets the same intervention as the experimental group, but later
    • Trying out a series of manipulations on the same group or in the same setting (e.g. the same school) over time