Designing and conducting observations

What are behaviour categories?

The behaviours you are tallying/observing

What does it mean to operationalise?

Break down behaviours into specific key actions

Why is it difficult to categorise behaviours?

You have to ensure no categories overlap and that there is a wide enough range of possible behaviours

What does sampling mean in the context of observation techniques?

The time period you observe

What is continuous recording?

When all instances of relevant behaviour are recorded, used in unstructured observations

What is event sampling?

The number of times a behaviour occurs during a particular event

Pros and cons of event sampling

Provides detailed frequency data of how often the target behaviour occurred

Simple to use

Could miss behaviours that occur at different times or in different contexts

Could be challenging to record complex behaviour or multiple simultaneously

What is time sampling?

Behaviour that occurs within a pre-established time frame

Pros and cons of time sampling

More representative, you can go back at various intervals

Reduces observer fatigue, shorter observation periods

Behaviours outside time intervals are missed

Not suitable for rare/infrequent behaviours

How do you measure the validity of an observation?

Face validity, ask another researcher if they believe the observation is valid

Interpretive validity, ensure events are interpreted the same way

How do you improve the validity of an observation?

Make the observation covert to increase validity

Ensure behaviour categories are clearly defined

How do you measure the reliability of an observation?

Inter-observer reliability, carry out a pilot study with 2 observers to check they interpret data the same way with a positive correlation of +0.80

How do you improve the reliability of an observation?

Properly operationalise behaviours so they are specific enough but don’t overlap

Video the event to go back and check your findings