Experimental Design: Example
The tree in the corner of of the field would tend to shade the plants growing close to it, which would reduce their performance compared to plants far from the corner. Without measuring plants of the same line in another plot, these plants would be classified as low yielding. Likewise, those planted near the edge of the plot may grow better because their roots have more space to exploit. Without randomisation, they would be planted on the field's edge in every replicate, and therefore appear to have better yield, which is an accident of where they were planted, and has nothing to do with their genetic make-up.