Time sampling in Structured Observation

This is also known as interval sampling because you observe for a specified duration at specified intervals – like taking a succession of snapshots, an analogy which is pretty close. Sampling means estimating the frequency of events in a continuous time sequence from a much shorter period (or periods) of time. The main practical challenge is deciding on the length of time for the sample estimate; there is a certain amount of trial and error in this.

The general principle is that the higher the frequency of the ‘events’ the shorter the periods of observation. There is a kind of gradient here: at some point, which can only be determined by the practicalities of recording as well as ‘representativeness’, the choice moves from event sampling to interval sampling (or the other way). With discontinuous interval sampling (as in taking a population sample for a survey) the level of representativeness needs to be gauged in some way.

Alternate sampling

Interval sampling has one important advantage in that you can alternate the focus of your observation. The focus could be:

• on some individuals in a group, e.g. boys as against girls in a classroom
• on some specific behaviours, e.g. how many children are not ‘actively learning’ in a classroom setting or more specifically are engaged in a particular learning activity – or not, as the case may be
• on different sites, e.g. in a public art gallery how many people are looking at one painting rather than another, or one kind of painting rather than another.

The question remains, however: how do you decide on the duration of the ‘interval’ in practice? Initially you should take a short interval – say every three minutes where you record for perhaps ten seconds – and then work out whether a longer interval (or a shorter duration of observation) would suffice. If the purpose of the research is the difference in gender response to teachers’ ques­tions then you might move the focus of recording from boys to girls (and back again) every other recording per­iod or you might cover all the children at the same time. Note that you would have to define in terms of specific behaviours what you mean by a ‘response’ (volunteering an answer, signalling the desire to do so, following work instructions).

And because children’s work in a classroom has a ‘devel­opmental’ sequence, in that it evolves throughout the les­son, then a consideration for sampling is whether these phases are adequately represented.

Source: Gillham Bill (2008), Observation Techniques: Structured to Unstructured, Continuum; Illustrated edition.