→ Random assignment
Is the difference between groups based on how they were formed or can we discount that as a reasonable explanation?
- Example: Psychological study whether participants were more creative after thinking of their intrinsic vs extrinsic motivation for writing
- Only calculating the mean fails to account for variability within each group
- Variability can be measured with the standard-deviation
If we assign participants to groups randomly, is it possible that an “unlucky random assignment” has an influence on the mean result?
- Example: Shuffle creativity scores and randomly assign to group, then calculate how often the mean creativity score for intrinsic and extrinsic motivation changes over, say, 1000 trial runs → New p-value