6.2 observation

Observation = a non-experimental technique. The researcher watches and records behaviour without manipulating the IV

Controlled observation = aspects of the environment are controlled. Often in a lab setting

Evaluation

  • reduces extraneous variables, so can establish cause and effect

  • Results are more reliable as they use standardised procedures

  • Artificial environment, So behaviour isn’t representative of real life

Naturalistic observation = takes place in the real world

Evaluation

  • high realism, participants are more likely to show naturalistic behaviours

  • High external validity, behaviour is generalisable to other situations

  • Lack of control over extraneous variables, so lower internal validity

Overt observation = participants are aware they are being observed

Evaluation

  • ethical as theres informed consent

  • Leads to demand characteristics or social desirability bias

Covert observation = unaware they’re being observed

Evaluation

  • more natural behaviour. No demand characteristics or social desirability bias

  • Unethical as they cannot give informed consent

Participant observation = the researcher joins the group being observed

Evaluation

  • can build rapport, so may disclose more info and behave more naturally

  • May lose objectivity, being biased (‘going native’)

Non participant = the researcher is separate from the group

Evaluation

  • will remain objective

  • Lack of trust/ rapport, misses out info and they don’t behave naturally

Observational design = the choice of behaviours to record and how they are measured

  • operationalised behavioural categories - the behaviours need to be measurable and identification

  • Time sampling - recording all relevant behaviour at set points, e.g every 15 seconds

  • Event sampling - records every time a behaviour occurs from the list of operationalised behavioural categories

Evaluation

Time

  • more flexibility to record unexpected behaviour

  • Can miss behaviour that happens outside of the recording periods

Event

  • Will identify the behaviours

  • May miss other relevant behaviours that isnt on the list of behavioural categories

Assessing reliability

Inter rater reliability

  • two or more observers conduct the same observation

  • They must use the same operationalised categories, conduct the observation separately, then compare the two and test correlation (e.g spearman’s rho)

  • A correlation of 0.8 or stronger is accepted