AtlasSession 3L7: Systematic errorsp.54–60, 86–87
Measurement of outcomes
- Study designs primer
- What is systematic error
- Measurement of outcomes
Fig. 1
Goat osteochondral defects: ICRS II histology score (0–100), seeded scaffold vs microfracture, true difference 0
Fig. 1: 30 sections per arm, seeded scaffold against microfracture, scored for ICRS II histology score; the true difference is 0. The scorer knows which group each slide is from; 9 borderline sections per arm glow. Seeded scaffold looks about 2 points better.
- Scorer knows the groups?
- yes
- Seeded scaffold looks better by
- about 2 points
- Real difference
- none
Agreement & Error Lab · First look
Level 1 · First look · Lecture 7 · p.54–60, 86–87
See how knowing which group is which can tilt how results are judged.
Name the three p.58 mechanisms (assessor aware of allocation, different methods per group, error related to the intervention) and predict how each skews a comparison.
Taste test where you know which drink is which (the step you are on)
Tasters who see the labels rate their favourite brand higher.
In a taste test where you can see the labels, your favourite brand tends to score higher. Here the scorer knows which slides came from the new scaffold: it looks better by about 2 points, though the real difference is none.