Observational Methods

• • Observational methods involve the observation of PP's behaviour- the researcher watches & interprets the behaviour.
  • They can be used in experiments to measure the dependent variable- but merely collecting the data by observation doesn't make it an experiment.
• Carrying out observations:

  1. Decide on the behaviour to be observed.
  2. Decide on the observation type- naturalistic vs controlled, participant vs non-participant, covert vs overt.
  3. Decide on a method to record data in observations- structure or unstructured.
  4. Choose behavioural categories- divide the behaviour interested in into subject of behaviours which can more easily be recorded. Variables must be in defined, measurable, distinguishable, observable categories.
  5. Choose a sampling method (need a systematic way of recording observations)- time sampling or event sampling.
• Naturalistic observations:

  • Observing naturally occurring behaviour in the setting in which it would naturally occur.
  • Can be used in quasi experiments (eg Darwin).
• Controlled observations:

  • Performing an observation within a structured or controlled environment.
  • Some variables are controlled so you can perform the observation in a lab (eg Bandura).
• Naturalistic Observations Evaluation:

  • High external validity- can generalise findings to everyday life.
  • Lack of control makes replication difficult- affecting the ability to check the reliability of results.
  • Lack of control also means the extraneous variables are more likely to have an impact on the results so we can say they have low internal validity.
• Controlled Observations Evaluation:

  • Low external validity- can't generalise findings to everyday life, as we have measured behaviour in an artificial setting.
  • Good control makes replication easier, allowing us to check the reliability of results.
  • Good control means extraneous variables are less likely to have an impact on the results, so we can say they have high internal validity.
• Participant Observations- researcher joins in with the group being observed.
• Non-participant Observations- researcher watches from outside the group (eg an Ofsted inspector observing a lesson).
• In-participant Observations Evaluation:

  • Research can get a real insight into the situation- can increase validity of results.
  • Helps avoid observer effects.
  • The researcher may become too involved in the experiment they're participating in- may "go native"- identifying too strongly with those they set out to research & loose objectivity in their data collection.
• Non-participant Observations Evaluation:

  • (Essentially the opposite to in-participant observations).
  • Researcher remains outside of the experiments enough to remain objective.
  • Does not have the same level of insight as in-participant observations- data may therefore be less valid.
• Covert Observations- participants don't know they're being observed, no consent is obtained so PPs are unaware.
• Overt Observations- participants are aware they are in a study & are being observed & have given informed consent. The observer would be clearly visible.
• Covert Observations Evaluation:

  • Have extremely good external validity- behaviour will be completely natural.
  • No participant reactivity.
  • Demand characteristics.
  • Investigator effects.
• Overt Observations Evaluation:

  • (Essentially opposite to covert observations).
  • More likely to have factors that reduce the validity such as PP reactivity.
  • Can ensure it is more ethical, as you can gain informed consent & therefore access to more private behaviours- such as those that occur in the home.
• Structured Observations:

  • You have a system of how to record the observations set before you start.
  • Done through either the use of behavioural categories or a sampling method.
  • Not free to write down anything else you deem as noteworthy.
  • Means you have a numerical value for types of behaviours.
• Unstructured Observations:

  • The observer just records everything they think is noteworthy.
  • May be difficult to decide what is relevant.
  • May not be quick enough in getting everything down.
  • May be drawn to just recording what is 'eye-catching'.
• Time sampling:

  • Observe behaviours within or at a specific time frame, eg first 10 mins in every hour.
  • Misses behaviour occurring between intervals- so events happening between intervals aren't recorded, so reduces validity.
• Event Sampling:

  • Recording a specific event each time it occurs in a time-frame, eg a fight.
  • If the behaviour is very frequent, it can be difficult to record accurately.
  • It can be difficult to remain focused & can miss 'events', especially if there's a lot of categories.
  • Good when behaviour is very frequent, if behaviour is less frequent, there are fewer individuals to observe.
• Observations General Evaluations:

  • Observing behaviour eliminates the opportunity to lie- so may get more valid results.
  • Versus questionnaires or interviews.
  • Observer bias- the researcher may only record what they expect to see, leading to low inter-rater reliability, observers must be trained in using the coding system.
  • Ethical issues of covert observations: deception, consent, privacy.
  • Demand characteristics/ social desirability in overt observations.