Across
- 3. [BLANK] trends
- 4. unobserved variable causing both treatment and outcome
- 6. the variable whose causal effect you want to identify
- 7. ? A trio tint, panel dropout
- 8. [BLANK] bias: non-random sorting into treatment
Down
- 1. fake treatment test
- 2. the F-stat on your excluded instrument tests this
- 4. ? An anagram of mic loper - follows the instrument.
- 5. ideal design