Experimental Design
Online survey randomized controlled trial (between subjects).
Participants are asked to imagine taking a university entrance test for the first time while considering admission to a selective and demanding academic program. They then imagine receiving a (randomly assigned) score X. The assigned score is always below but close to the admission cutoff for the desired program. Participants then report their hypothetical self-evaluations and behavioral intentions.
The score X is randomly assigned between 598 and 601. The experiment tests whether receiving a round-number score (600) affects self-evaluation. The main comparison is between participants who receive 600 and those who receive 599. As placebo tests, I also compare 601 versus 600 and 599 versus 598 to examine whether non-round scores produce any effects.
There are two main treatment arms:
- Arm A (treatment, round): score = 600
- Arm B (control, non-round): score = 599
This score appears as part of the scenario presented to each participant ("your entrance-test score is X").
The main causal estimand is the average treatment effect of being assigned score 600 versus 599 on self-evaluation and behavioral-intention outcomes: E[Y | X = 600] − E[Y | X = 599].
As placebo arms, participants may also be assigned scores 598 (Arm C) or 601 (Arm D). These allow tests of whether adjacent non-round scores generate effects. Formally, the placebo tests are: E[Y | X = 601] − E[Y | X = 600] and E[Y | X = 599] − E[Y | X = 598].