Daily 18 - Mar 30

Class Performance

Students: 93 | Mean: 2.63 | Median: 2.5 | SD: 0.86

Scores ranged from 0.5 to 4 out of 4 points.

Score Distribution

Performance by Question

Questions

Q1: Demographic Controls and Gender Wage Gap

Adding demographic controls explains a small fraction of the gender wage gap.

  • Writing “large” — Many students said “large,” but Column 3 shows demographics explain very little.
  • Confusing significance with explanatory power — Statistically significant controls don’t necessarily explain a large fraction.

Q2: Team+AI vs Individual+AI Speed Difference

Team+AI groups are 5 percentage points slower than Individual+AI groups.

  • Using “log time” or “hours” — Many wrote “0.048 log time.” The answer is 5 percentage points.
  • Writing “4.8” without units — Close value but must specify percentage points.
  • Writing in minutes or seconds — The table reports percentage point differences.

Q3: Small Class Equivalent Years of Experience

Being in a small class is equivalent to a teacher with 10 more years of experience.

  • Wrong values (1.5, 2, 3, 7, 14, 21) — Divide small class effect by the experience coefficient.
  • Close values (9, 9.5) — Approximately correct, received partial credit.

Q4: Randomization and the CIA

Randomization makes potential outcomes independent of treatment. Without an RCT, the CIA (Conditional Independence Assumption) must hold.

  • Missing CIA — Some wrote “independent” but omitted the second term.
  • Writing “CSA” or “selection” — The correct abbreviation is CIA.

Key Takeaways

Strengths: CIA concept well understood | Independence from randomization clear.

Review:

  • Demographics explain little of the gender wage gap
  • Read table units carefully — “5 percentage points” not “0.048 log time”
  • Small class ≈ 10 years experience — Divide class size effect by experience coefficient