Research Methods Flashcards
Hypothesis
prediction!
- Null = no difference
- Directional = positive/ negative change
- Non-directional = no previous research, so states difference but not the way it goes
IV
variable manipulated
DV
variable measured
CV
something other than the IV influencing the DV, varies along with the IV
EV
need to control these, affect the DV, becomes a CV when it varies along with factors interested in
Operationalisation
defining variables
Independent measures design
a separate group of Ps per condition of the study
:) - no order effects, no fatigue
:( - 2x as many people, individual differences
Repeated measures design
the same group of Ps to each condition of study
:) - no individual differences, fewer people, can be counterbalanced
:( - order effects, demand characteristics, fatigue
Matched pairs design
separate groups per condition of study, pairs matched on certain characteristics, like age, gender, IQ
:) - controls order effects, reduces individual differences
:(- takes more time n effect, still a chance of individual differences
Random sampling!
all members of target pop have an equal chance of being selected, like names in a hat
:) - most representative
:( - still not fully representative, time consuming, miss people, some may refuse to take part
Opportunity sample
use Ps available e.g. psychology class
:) - easy, less expensive, less time consuming
:( - very unrepresentative, sample bias
Volunteer sample
adverts. those who reply are in!
:) - reaches target populations
:( - volunteer bias = sample bias, those who want to take part may have similar characteristics
Systematic sample
choose every nth (e.g. every 3rd) member of target POP.
:) - no chance of researcher bias
:( - still possible to be unrepresentative
Stratified sample
use groups representative of target pop e.g. if target pop has 70% males and 30% females use this percentage.
:) - avoids researcher bias, very representative
:( - sample cant reflect all the ways people differ, we are complex and different
Lab experiment
controlled environment, IV manipulated by experimenter
:) - internal validity
:( - artificial, external validity
Field experiment
IV manipulated by experimenter, natural environment
:) - better external validity
:( - poorer internal validity, loss of control, EVs
Natural experiment
where IV naturally occurs
:) - high external validity, allows study of topics that are hard to control
:( - no random allocation = confounding variables , cant be replicated - low reliability
Quasi experiment
contain a naturally occurring IV the researcher examines the effects of on DV in controlled conditions
:) - high internal validity
:( - confounding variables, no random allocation
Primary data
first hand! specific to aim of study BUT harder to obtain
Secondary data
pre-existing, not specific to aim of study
Qualitative data
non- numerical
rich and detailed
harder to compare
Quantitative data
numerical
easy to compare
lacks detail
evaluate measures of central tendency
Median + mode- unaffected by outliers but dont take all scores into account
Mean- looks at all scores, affected by extreme scores
Range
largest- smallest
easy to calculate
influenced by extreme scores
SD
standard deviation - average distance from mean. the measure of dispersion of a set of data from its mean.
- takes all scores into account
- harder to calculate
Peer review
Research checked by other experts
:) - check quality of research,can make amendments, those with more good peer review likely to review funding for future research
:( - anonymity, publication bias (only significant research published) , new research
consent
-Ps must agree to take part in the study
HOW TO DEAL W IT: gain consent, debrief (confirm true aims of study) Presumptive consent (ask group similar)
Right to withdraw
-Ps must be allowed to leave at any time
HOW TO DEAL W IT: Debrief: allow Ps to remove data, even after the study
Deception
Ps should not be lied to
HOW TO DEAL W IT: debriefing
Confidentiality
Info should be kept private
-Do not publish names, addresses, pictures
Protection from harm
both psychological and physical
- offer sources of help
- remind them of right to withdraw
controlled observation
control over environment
:) - internal validity, control
:( - mundane realism, external validity
naturalistic observation
real world setting
:) - external/ ecological validity
:( - lack of replicability, poor control
participant observation
researcher gets involved e.g zimbardo as prison warden
:) - rich detail
:( - experimenter effects/ subjectivity
covert observation
undercover
:) - realistic behaviour
:( - poorer ethics
overt observation
out in the open
:) - more ethical
:( - demand characteristics
structured observation
a set method
:) - quantitative, easier to record / analyse data, can replicate
:( - external validity
Unstructured observation
natural e.g. park
:) - qualitative, richer detail
:( - observer bias
Behavioural categories
:) - objective
:( - hard to include all behaviours and distinguish between them
event sampling
recording particular behaviour
:) - if event is infrequent
:( - hard for complex events
time sampling
record particular time periods
:) - fewer observations
:( - unrepresentative
structured interview
:) - easy to replicate
:( - not flexible
unstructured interview
:) - can be flexible
:( - hard to replicate and compare, social desirability bias
questionnaire
pre-set written Qs
:) - easy to compare, quick
:( - social desirability bias
open self- report
:) - qualitative, rich detail
:( - hard to compare
closed self report
:) - quantitative, statistical analysis, replicable
:( - can’t expand
Correlations
Coefficient- strength & direction of a correlation
-1 to 1
closer to -1 or 1 = stronger negative/ positive C
scatter graph
for correlations, plotted points
evaluate correlations
:) - quick & economical, can lead to other research
:( - no cause & effect, third variables ignored
pilot study
small scale version of full study, test if it is good - wont waste money in real thing
bar chart
separate bars, discrete data
histogram
bars touch, continuous data
double blind trial
removes investigator effects
Ps and experimenter dont know the aim of the study
