Experimental and quasi experimental design Flashcards
what must a experimental design include
Randomization, Control, Manipulation
what is randomization?
when subjects are randomly assigned to intervention or control group
T/F: randomization increases selection bias
false
T/F: randomization increases validity
true
random sampling
picking random people to get into the study
random assignment
randomly put into groups after chosen
control
group that there is no intervention
T/F: a control group decreases bias
true
T/F: a control group decreases internal validity
false
manipulation
independent variable is manipulated
power analysis
test to see how many subjects are needed to obtain the desired effect size
types of experimental designs
randomized control trial, Solomon four group design, after-only
randomized control trial
sample, baseline data, randomization, intervention group, control group, postintervention data collected from both
Solomon four group design
sample is selected randomized and split into 4 groups. both have an 2 intervention groups and 2 control groups. 2 has no baseline data collected. all have postintervention data collected
after only experimental design
sample is selected, randomized into intervention and control group, data only collected at the end
advantages of experimental design
high level of evidence, can evaluate cause-effect relationship, minimize threat to internal and external validity, finding more likely to be generalizable
disadvantages of experimental design
complex, costly, sample size, difficult to implement in real world, ethical considerations
quasi-experimental design
no randomization and/or no control group or neither, has to have manipulation
non equivalent groups design
experimental group, control group, baseline data collected, intervention applied to experimental, data collected in the end
after only nonequivalent group design
experimental, control, no baseline, intervention to experimental, data collected after
one-group design
experimental, baseline, intervention, data collected
time-series
experimental data collected, data collected over time, intervention, data collected twice
advantages to quasi-experimental
practical, less costly, more generalizable
disadvantages of quasi-experimental
unable to distinguish casual relationships, lower level of evidence
