Observations

• Naturalistic observation

  Behaviour is left as it normally would be. Behaviour is observed within its natural environment.
• Controlled observation

  Some behaviour or aspects of the investigation are controlled by the researcher. Participants are now likely to know that they are being observed.
• Structured observation

  The researcher has a number of systems in place in order to record what they see. These systems may be behavioural categories event/time sampling procedures.
• Unstructured observation

  The researcher has no systems in place in order to record what they see. They simply record all that they see.
• Participant observation

  Involves the observer becoming actively involved in the activities of the people being studied. They observe from the participants POV.
• Non-participant observation

  The observer simply observes the behaviour of participants and does not get involved with the activities.
• Overt (disclosed) observation

  Observer's identity is known and the participants are aware of their purpose.
• Covert (undisclosed) observation

  The observer's identity is not known to the participants. Participants may be aware of their purpose but they do not know of their role as an observer; their purpose is hidden.
• When doing an observation, researchers need to decide how different behaviour should be categorised. These should be:

  • Objective: the observer should not have to make inferences about behaviours, but should just record explicit actions.
  • Cover all possible component behaviours and avoid a 'waste basket' case.
  • Be mutually exclusive: you should not have to mark two categories at any one time.
• Researchers create behavioural categories by:

  • Dividing a target behaviour into a subset of behaviours. You do this by breaking a stream of behaviour e.g. aggression into different behavioural categories e.g. shouting, hitting..
  • Each behavioural category needs to be operationalised e.g. upset cannot be measured but crying can.
  • The list of behavioural categories needs to cover all areas f the target behaviour e.g. no 'other' category is allowed.
  • The behavioural categories should not overlap.
• What categories might you use to assess whether a mother and infant are attatched?

  Infant cries when mother leaves, hand holding, infant rejects playing with stranger, hugs/hide with mum when stranger enters, willing to explore environment when mother is present, reunion mbehaviour.
• When conducting an observation the observer should record every instance of the behaviour in as much detail as possible. This works when the behaviour of interest doesn’t occur very often. But in many situations there would be too much data to record so a systematic technique must be used.
• Event sampling


  Counting the number of times a certain behaviour happens in a target
• Time sampling


  Recording behaviours at a given time
• Evaluation of Event Sampling


  +More valid/accurate/reliable

  +Less likely to miss predefined events

  -Easy to get overwhelmed

  -Can miss events not coded for
• Evaluation of Event Sampling


  +More structured, less likely to get overwhelmed

  +More representative over time

  -Could miss certain behaviours
• Observation studies in general

  +More ecologically valid

  -No control; less reliable

  -More subjective
• Controlled observation

  +Reliability; can be repeated

  +Control

  +Validity

  -Less ecological validity
• Naturalistic observation

  +Ecological validity

  +No demand characteristics

  -Ethical issues

  -Lacks control
• Overt observation

  +Ethically secure

  -Lacks ecological validity

  -Demand characteristics
• Covert observation

  +Ecological validity

  +Lacks demand characteristics

  +Control

  -Ethical issues
• Participant observation

  +Objectivity

  +Less ethical issues

  -Lacks ecological validity

  -Demand characteristics
• Non-participant observation

  +Objectivity

  -Unethical