what is reliability?
- consistency
- research can only be trusted if you can tell from the results from a study aren’t a one off or fluke
what is internal relibaility?
has the research been kept the same (standardised) for all participants such that it is replicable and could be repeated by another researcher to see if they get the sample results
what is external reliability?
has the research been carried out on multiple participants? for results to suggest a consistent trend, there needs to be a large sample of participants. if there are too few, then anomalies or “flukes” can skew the data
what is inter-rater reliability?
when there are multiple observers who watch the same participants and the compare their results to see how similar they were
features kept the same in milgram’s study
- confederate
- 15V interval
- sample shock
- makes noise/bangs wall at the same time
- prods
- word pair task
- participants is always the teacher
why would the procedure not have been completely standardised for all participants in milgram’s study?
- even though there were 4 prods, it would’ve been difficult for the experimenter to act in precisely the same way with each participant
why would milgram’s study have been (or would be) difficult to repeat?
ethical guidelines mean that this study couldn’t be carried out now - high levels of deception + likihood of harm being caused to participants
features kept the same in piliavin’s study
- gender of observers, victims, models
- outfit for victim
- train journey
- time of incident
- victim in critical area
why would the procedure not have been completely standardised for all participants in piliavin’s study?
- a passenger may have been standing in the place the ‘victim’ was supposed to collapse, meaning he couldnt always collapse in the same place
- floors of carriages may not have been equally clean (or dirty) across 103 trials
why would piliavin’s study have been (or would be) difficult to repeat?
- practical issues - e.g. finding a train with same time and no delays
features kept the same in bandura’s study
- physical aggression shown to doll
- verbal aggression shown to doll
- timing for each stage
- toys
- pre-test scenario
why would the procedure not have been completely standardised for all participants in bandura’s study?
adults cant act in precisely the same way with every child
also other factors such as home life
why would bandura’s study have been (or would be) difficult to repeat?
ethical guidelines mean this sudy couldn’t be carried out now - bandura did not follow a single guideline
was the sample size large enough in milgram’s study to see a consistent result and avoid flukes?
- no as 40 is quite small
- also yes, as they’re being observed individually which could still be good
was the sample size large enough in piliavin’s study to see a consistent result and avoid flukes?
yes - large number of people in a carriage for each trial
was the sample size large enough in bandura’s study to see a consistent result and avoid flukes?
no - not many children in each experimental group
