research methods Flashcards
type 1
false positive
incorrectly accepting alternate
significant but not
too high p- level
type 2
false negative
incorrectly accepting null
too low p level
not enough room for error
too strict
is actually significant
peer review AO1
- recommendations on whether should be published, whether needs revision
- validity
- appropriate methodology and procedure
- significant on human behaviour
- originality
peer review AO3
- anonymous, criticise rivals
bread jealousy - difficult to find suitable peer
especially for a new ground breaking topic
publication bias - prevent substandard research from entering
keeps reputation
less oppurting for plagiarism
observational techniques
covert
undisclosed
observing without knowledge
‘one way mirror’
- less likely investigator effects
- ethical issues
observational techniques
overt observation
open
know they are being observed
filming publicly
- more ethical
- investigator effects
observational techniques
participant observations
participant takes part
group member quietly observing without their knowledge (covert)
- in depth data, close proximity
- investigator effects if overt
observational techniques
non participant observations
doesn’t participate
quietly sits in corner
- less likely investigator effects becaude can’t see them
- lack of proximity
observational techniques
naturalistic
unaltered setting
doesn’t interfere in anyway
shopping centre
- ecological validity
- difficult for reliability because as you aren’t interfering so hard to be replicated
observational techniques
controlled observations
under strict conditions
lab setting, observation room
- relibaile
- low external validity
artificial environment
observational techniques
structured observations
coded schedules according to previously agreed formula organising data into behavioural catergorues
which behaviours examined
- compare across groups
- high internal validity
not full picture
observational techniques
unstructured observations
every behaviour recorded
much detail as possible
- richness in data
- observer bias
time sampling
records every behaviour between intervals
better use of time
not all behaviour relevant
event sampling
number of times behaviour occurs using tally
every behaviour
behaviour missrd if too much happening at same time