Observation Flashcards
OBSERVATION
research method used when watching participants directly to gather data about their behaviour
STRUCTURED
researcher uses behavioural checklist - predetermined list of limited behaviours, training for consistence
+ data collection is consistent/clearly defined
+ high inter-observer reliability
+ easy to analyse results - quantitative
- identifying behavioural checklist is time consuming
- may miss important behaviours if checklist is followed
NON-STRUCTURED
observer records all behaviours relevant to study’s aims - limited to pilot study to help set the focus - qualitative data
+ all important behaviours recorded
+ lots of rich data gathered
- difficult to record all data accurately
- may gather irrelevant data
- may miss important behaviours
COVERT
Ps don’t know that they’re being observed - hidden, CCTV
+ Ps interacts with observer in more natural way
+ lower risk of demand characteristics & social desirability bias
- ethical issues - deception, no consent
- data recording may be more difficult
OVERT
Ps know they’re being observed - explicitly told
+ more ethical - valid informed consent, no deception
- observer presence likely to alter behaviour of Ps - lowers validity of results
PARTICIPANT
researcher joins Ps as part of social setting - can be covert (undercover) or overt
+ produces ecologically valid, in depth, and detailed data - cuz researcher is closer to situation
- difficult to carry out - being part of social setting is distracting
- researcher presence increases demand characteristics
- observer may become less objective
NON-PARTICIPANT
researcher observes from a distance - can be covert (one way glass) or overt
+ observer more objective - less observer bias
+ increase accuracy - more focused on recording data
- may miss behaviours cuz they’re recording from a distance
NATURALISTIC
Ps observed in uncontrolled/natural/real world setting - no interference from researcher
+ increase eco validity
- no guarantee that target behaviours will occur spontaneously
- lack of control - increase uncontrolled variables = less reliable + difficult to replicate
CONTROLLED
environment manipulated socially/physically by researcher - mainly conducted in standardised/laboratory setting
+ interference may be neccessary
+ increase control
+ standardised
+ increase reliability and replicability
- researcher interference can lead to biased results
- low eco validity
BEHAVIOURAL CHECKLIST
researcher systematically classifies behaviours into distinct categories - coding using numbers
TIME SAMPLING
behaviours only recorded during specific time periods (e.g every 15 seconds)
EVENT SAMPLING
all occurences of target behaviour are recorded each time they happen
WHY DO WE USE OBSERVATIONS
- sometimes IV can’t be manipulated
- may be the most valid measure of DV
- cheap to carry out - few resources needed
- increase eco validity for some types of observations
TECHNIQUES TO KEEP VALIDITY HIGH IN OBSERVATIONAL RESEARCH
- double blind technique - both researcher and Ps are unaware of what the hypothesis is
- covert (one way mirror) - lowers demand characteristics, Ps less likely to alter behaviour
- clear definitions of behaviours to be observed - lowers observer bias
- inter observer reliability - ensure more than one observer and they compare results
- pilot study - practice run to see if all observers are observing in the same way
