AtlasSession 3L9: Variations in study designp.7–20, 40–41
Factorial trials and interactions
- 3 earlier topics
- Types of distribution
- Normal vs non-normal
- Choosing a statistical test
- Parallel and multi-arm trials
- Factorial trials and interactions
Fig. 1
Open tibial fractures, non-union at 12 months (%): rhBMP-2 (yes / no) × MSC injection (yes / no)
Fig. 1. Four groups of 75: rhBMP-2 or not, crossed with MSCs or not, 300 patients in all.
- Patients in all
- 300
- In each of the 4 groups
- 75
- Each question
- 150 with vs 150 without
Trial Design Board · First look
Level 1 · First look · Lecture 9 · p.7–20, 40–41
See how one trial can answer two questions by testing combinations.
Explain why a 2×2 factorial answers two questions with fewer participants, and classify an interaction as none, synergistic, antagonistic or negative.
Two-for-one tasting (the step you are on)
Sauce yes/no and spice yes/no on four plates; every plate informs both questions.
Like four tasting plates, sauce or not crossed with spice or not, where every plate answers both questions: 75 patients in each of the 4 groups, 300 in all.