All simulators
The Science Table
Twelve workers with both potential outcomes: hide the missing half, randomise, and watch estimates centre on the ATE.
Day 31: The Rubin Causal Model & Potential Outcomes| worker | T | Y(0) | Y(1) | effect | observed Y |
|---|---|---|---|---|---|
| A | 1 | ? | 14 | ? | 14 |
| B | 0 | 13 | ? | ? | 13 |
| C | 1 | ? | 8 | ? | 8 |
| D | 0 | 11 | ? | ? | 11 |
| E | 1 | ? | 9 | ? | 9 |
| F | 0 | 16 | ? | ? | 16 |
| G | 0 | 10 | ? | ? | 10 |
| H | 0 | 8 | ? | ? | 8 |
| I | 1 | ? | 13 | ? | 13 |
| J | 1 | ? | 13 | ? | 13 |
| K | 1 | ? | 11 | ? | 11 |
| L | 0 | 11 | ? | ? | 11 |
true ATE
1.17
hidden in practice
true ATT
1.50
treated workers only
difference in means
-0.17
what the data show
0 experiments; average estimate — versus ATE 1.167.
Twelve workers' earnings without (Y0) and with (Y1) training, in $1,000. In real data only one column per worker is visible; the other is '?'. Random assignment gives estimates that scatter around the ATE; self-selection (enrol if the gain exceeds 2) does not.