We are aiming to capture something which is usally not observable in a direct sense and if the capturing process is not good it will introduce a bias. A difference between the TRUTH and the estimation.
| Imprecise | Precise | |
| Biased |
|
![]() |
| Unbiased |
|
|
Doing comparisons is actually a way around bias, if we have a clock which is fast and we clock a dropping oject, we will see the time to be too long. However, we can still compare the objects dropping and do away with the absolute times. Useful images of this include weighing procedures, the scale is off but comparing things makes that disappear.
Bias in statistics is most often pictured as an off-target shooter, one who has a tropism.