Sample size (or number of clusters) by treatment arms
Student profile attributes are randomized within teachers.
Each teacher evaluates seven profiles in which the randomized attributes vary across profiles. Because attributes are independently randomized, treatment arms are defined by combinations of profile attributes rather than by assignment of teachers to treatment groups.
Across the full sample, observations will be approximately evenly distributed across the possible combinations of randomized attributes. For example, student gender has two arms (male, female), and approximately half of the profile evaluations are expected to be assigned to each.