observational designs + techniques

• types of observational designs:

  • nautralistic/controlled
  • covert/overt
  • participant/non participant
• naturalistic observational design:

  • takes place in setting where target behaviour would usually occur
  • all aspects of environment are free to vary
• controlled observational design:

  • structured environment
  • some control over variables - manipulation of variables to observe effects + control of CVs + EVs
• naturalistic strengths + weaknesses:

  • high external validity
  • lack of control makes replication difficult
  • uncontrolled EVs + CVs make it difficult to judge patterns of behaviour
• controlled strengths + weaknesses:

  • higher control of EVs + CVs
  • replication becomes easier
  • low external validity
• covert observational design:

  • participants unaware they are being observed
  • consent has not been obtained, behaviour must be already happening
• overt observational design:

  • aware they are being observed, given informed consent
• covert strengths + weaknesses:

  • removes demand characteristics
  • high internal validity so behaviour is natural
  • ethics - privacy + lack of consent
• overt strengths + weaknesses:

  • more ethical

  
  

  • more likely to change behaviour due to known observation
• participant observational design:

  • researcher becomes part of group theyre studying
• non participant observational design:

  • researcher remains separate
  • recorded more objectively
  • may be impractical/impossible to join certain groups
• participants strengths + weaknesses:

  • can experience situation - greater insight
  • increased external validity

  
  

  • may lose objectivity if they identify strongly to people they are studying
• non participant strengths + weaknesses:

  • objective + psychological distance
  • less likely to affect group dynamic
  • may lose valuable insight - too far removed
• limitations of observational method (observer bias):

  • interpretations affected by expectations
  • issue when there is mainly one researcher - using >1 observer improves reliability
  • assessed through inter rate reliability
• limitations of observational method (cant demonstrate casual relationships):

  • not directly measuring effect of IV on DV
• observational design:

  • unstructured/structured
  • behavioural categories
  • sampling methods
• unstructured design:

  • writes everything happening
  • appropriate when small scale observation conducted
• structured design:

  • target behaviours are simplified
  • use of behavioural categories + sampling methods
• behavioural categories:

  • target behaviour broken up into components that are observable + measurable (operationalisation)
  • all observers need to interpret same behavioural category in same way - must be operationalised, objective + not overlapping
  • before observation =researchers must include ways in which target behaviour may occur (behaviour checklist)
• sampling in observational studies:

  • unstructured: continuous recording
  • structured: systematic way of sampling
• event sampling:

  • researcher records each time target behaviour occurs
  • useful when target behaviour occurs infrequently
• time sampling:

  • records behaviour at a fixed time frame - decreases amount of observations made
  • instances when behaviour is sampled may be unrepresentative of observation as a whole
• unstructured strengths + weaknesses:

  • rich detailed data - qualitative
  • greater risk of observer bias - what catches eye
  • unsuitable for large samples
• structured strengths+ weaknesses:

  • recording is easier + more systematic
  • quantitative data - analysis is easier

  
  

  • less detail