AtlasFoundations · start here
Comparing groups: real difference or chance?
- What is data?
- Averages and spread, by eye
- Samples and populations
- Comparing groups: real difference or chance?
Fig. 1
Cell attachment at 24 h on plasma-treated vs untreated PCL films (cells/mm²)
Fig. 1. 6 films per group: untreated averages 1,180 cells/mm², plasma-treated 1,450 cells/mm², a gap of 270 cells/mm². The spread within groups is 150 cells/mm².
- Untreated average
- 1,180 cells/mm²
- Plasma-treated average
- 1,450 cells/mm²
- Observed gap
- 270 cells/mm²
Sampling Simulator · First look
Level 1 · First look · Foundations (ext)
See that two groups almost always differ a bit, and ask whether this gap is more than luck.
Judge informally whether a gap between two groups is bigger than chance usually makes, and explain why that question needs a test.
ext A Foundations primer: written for beginners, not from the slides.
Two coins, two scores (the step you are on)
You get 4 heads in 10, your friend gets 6; both coins are fair (cf. L5 p.62).
Two fair coins can give 4 and 6 heads in 10 tosses: two groups almost always differ a bit, by luck alone. These two groups of 6 films differ by 270 cells per square millimetre.