|
Field
Last Published
|
Before
December 02, 2020 10:59 AM
|
After
December 02, 2020 11:57 AM
|
|
Field
Planned Number of Clusters
|
Before
54 discussion sections.
|
After
109 discussion sections.
|
|
Field
Planned Number of Observations
|
Before
1,000 students in data collection phase and 300 students in the experimental phase. Actual numbers depend on fractions of students who consent to participate and fraction scoring sufficiently low on the first exam.
|
After
2,000 students in data collection phase and 600 students in the experimental phase. Actual numbers depend on fractions of students who consent to participate and fraction scoring sufficiently low on the first exam.
|
|
Field
Sample size (or number of clusters) by treatment arms
|
Before
100 in each treatment arm and 100 in control for a total of 300.
|
After
200 in each treatment arm and 200 in control for a total of 600.
|
|
Field
Power calculation: Minimum Detectable Effect Size for Main Outcomes
|
Before
With standard errors clustered at discussion section level, we can detect an effect of 4 percentage points for overall course score 80 percent of the time. This assumes a standard deviation of 9 in overall course scores. We can detect an effect in office hour visits of 0.75, with standard deviation of 1.7. We can detect an effect of 6 percentage points on the second exam, with standard deviation 14.5.
|
After
With standard errors clustered at discussion section level, we can detect an effect of 2.8 percentage points for overall course score 80 percent of the time. This assumes a standard deviation of 9 in overall course scores. We can detect an effect in office hour visits of 0.53, with standard deviation of 1.7. We can detect an effect of 4.2 percentage points on the second exam, with standard deviation 14.5.
|