Validity, Reliability and Observations

• external validity


  concerns whether the results can be generalised
• population validity


  concerns whether the results can be generalised to other people
• temporal validity


  concerns whether the results can be generalised to other times
• ecological validity


  concerns whether the results can be generalised to other settings
• internal validity


  Concerned with what happens within a study - the extent to which a researcher is measuring what was intended
• demand characteristics


  When a participant picks up cues during an experiment that unconsciously makes them aware of the aims of the study and modifies their behaviour. Known as the “please you or screw you” effect.
• investigator effects 


  Anything the investigator does that has an effect on a participants performance in a study other than what was intended. This could be direct (interacting with participants) or indirect (as a result of the investigator designing the study).
• mundane realism

  Whether or not the task the participants are asked to do represent something

  they would experience and do in real life. 
• controlled observations


  This is when behaviour is observed in a controlled setting, but under conditions where certain variables have been organised by the researcher.
• naturalistics

  
  

  This is when an observation takes place in an everyday setting, in which the investigator does not interfere and merely observes the behaviour in question.
• overt

  This is when the participant is aware that their behaviour is being studied.
• covert 


  This is when the observation is kept secret and the participants do not know that they are being observed.
• participants observations 


  Observations are being made by someone who is  also participating in the activity being observed.
• non participant


  The observer is separate from the people being observed.
• What is the main characteristic of unstructured observations?


  All behaviour that is seen is recorded.
• What type of data do unstructured observations produce?


  Qualitative data
• How many participants are typically involved in unstructured observations?


  Few participants
• What is a key feature of structured observations?


  The observer uses systems to organize observations.
• What type of data do structured observations produce?

  Quantitative data
• What is the scale of participants typically involved in structured observations?


  Lots of participants
• What are the steps involved in creating behavioral categories for structured observations?


  • Break target behaviour into observable categories
  • Ensure categories are clear and unambiguous
  • Include all possible forms of target behaviour
  • Avoid 'dustbin categories'
  • Ensure categories are exclusive and do not overlap
• Why is it important for behavioural categories to be observable and measurable?


  To ensure that they do not require further interpretation.
• What could happen if behavioural categories overlap?


  It would be difficult to discern between different behaviours.
• What is time sampling in structured observations?


  Recording behaviour categories within a pre-established time frame.
• What is a strength of time sampling?


  It reduces the number of observations that have to be made.
• What is a limitation of time sampling?


  Behaviour sampled might be unrepresentative of the observation as a whole.