Yield of tomato plants
Signal = difference in yield caused by fertiliser
Noise = natural variability in yield
The simulation is initially set up to make the fertiliser increase yield by 0.5 kg per plant. Mention that the two distributions overlap. The noise makes it hard to assess the effect of the fertiliser (or even whether it has any effect) from the 20 experimental yields.
Click Vary effect and use the slider to investigate how much improvement must come from the fertiliser in order for us to be sure from the experimental results that fertiliser does improve yield.
Mention that statistics tries to isolate the signal from the noise in data like this.