Other Considerations

Misconception about data

When interpreting our plots, we should always seek to validate what we see in the context of the real world. If, for instance, the data is too good to be true, then it probably isn’t true.

Example

For example, residual plots, the points that don’t fall out right next to the regression line, should generally contain a visually random element. They should, in other words, present no visible pattern. That would suggest a bias in the data.