GRAVE:
generalisability- all studies involved large samples, anomalies should cancelled out, but all children came from the Stanford uni nursery. also difficult to generalise to adults who are less influenced by role models.
reliability- was replicated 2 times by himself, as the procedure was standardised. having two observers behind the mirror rules our subjectivity andtherefore inreases the interrater reliability
applications- 65, media censorship, showing violoence to young children might have an effect on their real life, violence on tv after 9pm.
validity- the children were put into an unusual situation, where they were shown unusual behaviours, and maybe they thought the experimenters wanted them to display the same behaviours, and therefore there is a problem with demand characteristics.
ethics- the children in the study may have been distressed by the harmful and violent behaviours, and they behaviours which they learnt might stay with them creating behavioural issues.