Research Methods Flashcards
Operationalisation
Turning a construct into an operational variable with an objective measure
Reliability
the consistency and stability of a measure
Validity
The measure measures the construct it’s supposed to measure
Test-Retest Reliability
giving a participant the same test multiple times
Interrater Reliability
Agreement amongst raters observing behaviour
Internal Reliability
e.g. through split half reliability which splits items in measure randomly into two and calculates the scores for each half –> checks relationships between the two halves
Cronbach’s alpha
Average of all possible split half reliability scores
Internal validity
validity within the experiment
external validity
validity outwith the experiment, e.g. ecological, population, temporal
Face validity
intuitive sense of the experiment
content validity
the measure reflects the conceptual definition of the construct from the literature represented in the experiment
Criterion validity
the extent to which people’s scores on a measure are correlated with other variables that one would expect to be correlated with
including concurrent validity and predictive validity
concurrent validity
criterion and construct are measured simultaneously
predictive validity
the ability of a test or other measurement to predict a future outcome
convergent validity
correlates strongly with the measure of the same construct, different tests that measure the same thing
discriminant validity
correlates less with the measure of a different construct
Within Subjects design
All participants are exposed to all levels of the IV at different times
Between subjects design
Different participants are exposed to different levels of the IV
extraneous variables
any variable other than the DV or IV that could affect your outcomes
confounding variable
an extraneous variable that varies systematically across levels of the IV
Block randomisation
All the conditions occur once in the sequence before any of them is repeated
Complete counterbalancing
equal number of participants complete each possible order of conditions
Randomised control trial
experiments done in the field to investigate the effectiveness of a treatment/intervention
random counterbalancing: Latin square
each condition appears exactly once in each position
Placebo effect
the mere expectation that there will be an improvement contributes to said improvement
Internal validity
Is there a causal relationship between the variables - quality of the research design
statistical validity
proper treatment of the data and the soundness of the researchers statistical conclusions
external validity
generalisability of the findings
mundane realism
participants and situation studied are similar to those the researchers want to generalise to
psychological realism
same mental process is used in both the lab and the real world
single factor two level design
studies the effect of only one IV on the response of a DV
single factor multi level design
studies the effect of multiple IVs on the response of a DV
Non-experimental design
looking at associations between variables with no random assignment, manipulation of IV or control group
Observational Research
Systematic observations in a non-experimental design
Naturalistic observation
field work, in the wild research
disguised: participants are unaware of being studied -> avoids experimenter effects, higher validity but ethical issues
undisguised: participants aware of researcher’s presence -> reactivity - different due to observation? more ethical?
Participant observation
becoming part of the participants
disguised: infiltration of group
undisguised: disclosure of identity
can be inorganic due to researcher’s influence on group, and group’s influence on researcher,
high ecological validity, time consuming, ethical implications
Structured/controlled observation
Observations made in an artificial environment to observe specific behaviours
Behavioural coding
practice of formally and systematically defining overt (observable) behaviours
time sampling
method of collecting data in which you watch research participants for a specific amount of time at specific time intervals and record behaviour
event sampling
record each time a certain behaviour occurs
Quasi experimental Research
involves manipulation of IV but no random assignment of groups
Post test only design
IV is given to group and DV is measured afterwards
Pre-test-post-test design
DV is measured before and after treatment
Regression to the mean
extreme variables tend to level out if repeatedly measured
statistical phenomenon that makes natural variation in repeated data look like real change
Non equivalent pre-test post-test design
different groups get different conditions
interrupted time series design
data is measured at multiple time points before and after the introduction of an intervention to examine effect of the intervention
Survey
Involves design, data collection (questionnaire or interview) and analysis and interpretation
can be experimental or non-experimental, non-experimental most of the time
Stages of Survey construction
Aims and objectives of the survey, target group
Question style
Question Design (Brief Relevant Unambiguous Specific Objective)
Double Barrelled questions
A question that asks more than one questions
e.g have you experienced depression AND anxiety in the last 2 weeks?
Leading Questions
A question that prompts or encourages a specific answer
Acquiescence Bias
Yay-saying (always agreeing)
Nay-saying (always disagreeing)
fence-sitting (always picking middle category)
-> fix through reverse scoring items, even numbered rating scale
Piloting
testing an experiment on a small target population
representative sample
a subset of people that portray similar characteristics of the population that you are aiming to investigate
stratified random sampling
A population is divided into homogenous subpopulations, or strata, based on specific characteristics (race, gender, identity)
each group is randomly sample
simple random sampling
every member of a population has an equal chance of being selected -> representative and unbiased
-> needs info of entire population
cluster sampling
a population is divided into smaller groups or clusters and some clusters are selected randomly. used to study large population, usually pre-existing clusters (e.g. schools/cities)
can be biased if clusters aren’t representative
quota sampling
sample represents estimated numerical composition population - not randomly selected
recruit until quota for subgroup is reached
systematic sampling
select members of the population at a regular interval (k) determined in advance. time efficient, fixed interval reduces bias, not recommended for periodically ordered groups
snowball sampling
start with a few participants and then continue to recruit people based on referrals from initial participants
convenience sampling
sampling those who are easy to contact, volunteers
purposive sampling
obtaining a sample with pre determined criteria, not meant to make statistical inference to whole population