NURS 231 Lecture Notes - Lecture 6: Simple Random Sample, Stratified Sampling, Pilot Experiment
231 module 6
3 criteria for an experimental design
• Randomly designed which makes both groups even
• There is an intervention
• There is a control group
Solomon four group experiment
Group a pretest yes intervention yes post test yes
Group b pretest no intervention yes post test yes
Group c pretest yes intervention no post test yes
Group d pretest no intervention no post test yes
Non-experimental designs
Still looks at causes but no experiments because sometimes its not ethical to do n
experiment.
Correlation doesn’t mean cause. So correlational studies cannot be looked at as
cause and affect it’s more a relationship. To have cause you need a study and you
need an intervention. So surveys cannot have cause and effect.
Prospective is better than retrospective because in retrospective people cn forget.
Longitudinal is better.
Case control is finding a control group that is similar with the case group and
compare them. (retrospective) – these don’t show patter of change
Cohort studies means studying people over time (prospective)
Pilot study isn’t reliable information yet. Its finding out if the information is best to
carry out a bigger study. The information is good to note but isn’t reliable yet.
Determine the non experimental design most appropriate to use:
• 1. Correlational
• 2. Descriptive
• 3. Cross sectional
• 4. Cohort
• 5. Correlational
internal and external validity
internal affects external
history threat means there might have been a bias that affects the results of the
study. (internal)
find more resources at oneclass.com
find more resources at oneclass.com
Document Summary
3 criteria for an experimental design: randomly designed which makes both groups even, there is an intervention, there is a control group. Group a pretest yes intervention yes post test yes. Group b pretest no intervention yes post test yes. Group c pretest yes intervention no post test yes. Group d pretest no intervention no post test yes. Still looks at causes but no experiments because sometimes its not ethical to do n experiment. need an intervention. Prospective is better than retrospective because in retrospective people cn forget. Case control is finding a control group that is similar with the case group and. So correlational studies cannot be looked at as cause and affect it"s more a relationship. To have cause you need a study and you compare them. (retrospective) these don"t show patter of change. Its finding out if the information is best to carry out a bigger study. The information is good to note but isn"t reliable yet.