Some differences between the values of a variable are unavoidable, but we sometimes intentionally make changes to some individuals to observe their effect on the variables. The resulting variation is caused by the experiment.

In all cases, the noise in the data will make it harder to detect the signal.

An important role of statistics is to estimate the signal from data that also contain noise.


Effect of fertiliser on tomato plants

Note that the data from this experiment would be entered into a data matrix as shown below for analysis.

   Yield    Group
6.57
6.53
4.71
5.32
6.15
5.08
6.17
4.93
3.16
4.57
6.13
7.49
6.00
4.84
9.01
6.11
5.55
5.48
6.84
6.18
   No fertiliser   
No fertiliser
No fertiliser
No fertiliser
No fertiliser
No fertiliser
No fertiliser
No fertiliser
No fertiliser
No fertiliser
Fertiliser
Fertiliser
Fertiliser
Fertiliser
Fertiliser
Fertiliser
Fertiliser
Fertiliser
Fertiliser
Fertiliser