More and more, the omnipresent variation between individual items in a batch is taken into account by using special analysing techniques like mixed effects and indexed regression. In this paper the assumptions upon which these techniques are based, are explained, based on a simulated data set. The most important issue is the separation of biological variation and the technical variation (or measuring error). The techniques are further elucidated with some examples from practise (skin colour apples in storage, softening of melons in storage and water loss in plums, melons and mandarins), to show the increased reliability of the analyses. Since variation is always present in any measured data set, the techniques are applicable in all fields working with living material.