Abstract: All observations are a selection of all the animal behaviour that could be observed. This chapter discusses what this means in terms of designing a research project. Selection has to be made about between the following: (i) the level at which the observations will be made (group-individual animal- body part); (ii) the units (behaviour patterns) to be recorded; (iii) the type of sampling (focal animal, scan, ad libitum, behaviour) that will used; and (iv) the type of record (continuous, zero/one, or something in between) with the consequent type of measurement to be extracted (categorical, ordered (ranked), or interval).

Keywords: units of behaviour, levels of observation, sampling, focal animal sampling, scan sampling,