CH 10 - Non and Quasi Experimental Designs Flashcards
Non- & Quasi- Experimental Designs
For when a researcher is not able to manipulate an IV and must work with pre-existing levels of the IV.
EX: Effects of maternal smoking on infant birth weight - not random assignment
Or for when it is not necessary to manipulate the IV because information already exists,
EX: effects of speed limit reductions on accidents (just do before and after comparison)
Similarities & Differences
Similarities
- no manipulation of IV
- no randomization
Differences
- Is there an attempt to control?
Types of Between Groups Non- & Quasi- Experimental Designs
No assignment to groups = non-equivalent groups and potentially assignment bias.
Groups may differ on a variable other than your IV and this can be a possible confound and a threat to internal validity.
- Differential design N
- Post-test only NECG design N
- Pre-test/Post-test NECG design Q
Types of Within Groups Non- & Quasi- Experimental Designs
Pre-post designs
Comparing individual scores rather than groups
Assignment bias is eliminated
**However, you cannot counterbalance the order of treatments like in experimental design
- One group pretest-posttest design N
- Time series design Q
Threats to internal validity:
- History
- Maturation
- Instrumentation
- Order effects
- Regression to the mean
Differential design N - Ex Post Facto design
“After the fact”
Easiest of the 3 between-group designs
Compare pre-existing groups to find a difference on variable of interest.
Ex: Group 1 - Public CEGEP
Group 2 - Private
Correlational = “How does A relate to B”
Differential = “How does A differ from B”
Post-test only NECG design N
Examines effectiveness of treatment or intervention
No manipulation, no control over who is exposed
Compare those who received treatment vs those who did not
Pre-test/Post-test NECG design Q
Adds a bit of control by including a pre-test measurement for both groups.
Determine similarity in groups prior to administering treatment. Then if there is a different after treatment you can be more confident.
It reduces but does NOT eliminate assignment bias. You are controlling for one variable where with randomization you target multiple variables.
One group pretest-posttest design N
Two observations are made for each participant, one before and one after treatment.
If there’s a change post-treatment, we cannot be sure that it is due to the treatment and not some other factor.
Time series design Q
Adds some control by making a series of observations for each participant, some before and some after treatment.
Multiple observations allow you to see trends in your data and the effect of your treatment.
Developmental Research Designs
Non-experimental research design used to study changes in behaviour related to age
Cross-sectional DRD
Longitudinal DRD
Cross-Sectional Research Design
Between-groups where different age groups are compared on a variable of interest.
EX: differences in ease with technology or attitude towards public health behaviours.
Adv:
- time-efficient (don’t need to wait years for people to age)
- no long term cooperation required
Dis:
- cohort or generation effects (differences between groups attributed to characteristics or experiences other than age)
- individual changes not assessed
Longitudinal Research Designs
Within-group design that measures a variable n the same group over time.
Adv:
- no cohort or generation effects
- individual changes assessed
Dis:
- time-consuming
- participant attrition
- practice effects?
Corss-Sectional Longitudinal Research Designs - Mixed Developmental Designs
Compare the results obtained from separate samples (cross-sectional) that were obtained at different times (longitudinal).
Allows for the examination of the development of phenomena other than individual aging.
