research planning and design lecture 3 Flashcards
non experimental methods
non-experimental research design
- establishes association between variables
- no random assignment
- no manipulation of independent variable
- little to no control over extraneous variables
types of non experimental research design
correlational
observational
qualitative
observational research
= recording systemic observations of behaviour in a natural or laboratory setting without manipulation.
types:
naturalistic
participant
structured
naturalistic observation (field work)
= observing peoples behaviour in the environment in which it typically occurs
disguised -> participants are not aware they are being studied -> higher validity, avoids experimenter effects, but is it ethical? in a private or public space?
undisguised -> participants are aware of researcher and that they are being monitored, more ethical but will the participant act different?
participant observation
= researchers become active participants in the group or sitaution that they are studying
disguised -> pretends, conceal their identity
undisguised -> disloses their identity
benefits -> high ecological validity, true insight
costs -> time consuming, invasion of privacy?, bias due to getting used to and sympathising with group
structured observation
= observation made in an artificial environment to observe specific behaviours
behavioural coding:
time sampling = recording observations at specific time intervals -> reducing observations but might not be representative
event sampling = recording every time a behaviour occurs -> can miss behaviours if too many occur at same time
quasi experiments
- involves manipulating the independent variable
- cannot randomly assign participants to groups
- cannot control for confounding variables
- 3 types
one group post-test only design
= only a single group and testing is only done once, after the issue being studied has already occurred
very weak as there is no baseline, no comparison group
however it is quick to implement, good as an initial exploration
one group pre-test- post-test design
= the dependent variable is measured before and after the treatment
one group is bad design, practice effects, history, spontaneous remission, regression to the mean
regression to the mean
= makes natural variation in repeated data look like real change
- Individuals who scores extremely high or low on one occasion tend to score less extreme on the next occasion -> regress back to typical performance
non equivalent pretest-post-test design
= a treatment group is given a pre test, then receives a treatment, followed by a post-test. There is then a non equivelent control group that is given same as above but NO treatment.
