Ethical considerations in experimental design

Ethical considerations in experimental design
- Randomized clinical trials use one control and at least one experimental group to test
effects of a new medication, a therapeutic approach, or apparatus as treatment for a
medical or mental illness
- Research studies that test efficacy of new drugs pose risks to p’s – FDA requires new
drugs undergo several stages of testing, beginning w lab tests, etc. – takes a long time
o Bc of lengthy processes, argue for modified experimental design to eval benefits
of new and potentially lifesaving drugs
o Ex. Health advocates who challenged use of control groups in pre-posttest
studies of new HIV/AIDS drugs contend control group p’s experience a delay in
receiving potentially lifesaving drug
- Researchers must ensure that the new drugs r effective, r safe, and don’t cause life-
threatening complications; time needed to complete all testing procedures delays
distribution of potentially lifesaving drugs
- One way to address ethical dilemma thro a pre-post-post-test design – both exp and
control groups would receive actual treatment, not at the same time
- Experimental and control get the drug = intervention – experimental A group and exp B
get pretest at the same time
o Exp group gets intervention drug first, control is delayed
o Second posttest admin after delayed intervention
- By using pre-post-posttest design, ensure both groups get drug, but w minimal delay in
administering drug to control group, researchers can still conduct drug vs no drug
comparison to test drug’s safety and effectiveness
Quasi-exp intervention studies
- Adjusts for fact that researchers can’t always control each independent v
- Inability to control some IV means that researchers aren’t able to demonstrate a direct
cause-and-effect rel btw IV and DV or outcomes
- Don’t include control groups – may b unnecessary if researcher interested only in impact
of an intervention on study p’s – p’s r assigned to same condition
- Without control, researcher loses comparison of outcome of intervention to another
group – can’t claim cause-effect or generalize to larger pop’n
Table 2.2
Nonexperimental
- Qualitative: explore phenomenon in context; minimal or no stat data; in depth analysis
of response, but can’t examine cause – effect rels
- Case: in-depth exploration of person, place, situation; minimal or no statistical data; in-
depth exploration of rare/unique events, can’t examine cause-effect rel

- Focus groups: gather info, generate insight, explore decision making; minimal use of
descriptive data; generate new info, interactive approach, can’t examine cause-effect rel
- Corr studies: describe rel btw 2 v’s; pearson corr coefficient Range = -1.00 to +1.00; ID’s
rel btw 2 variables, can’t determine causal rel
Experimental
- Exp: detect cause-effect rel btw v’s; central tendency, student’s t, analysis of variance,
causal explanation of effects of one or more v’s on outcomes, direct control of causal v’s,
not suitable of all studies
- Intervention studies: measure effect of intervention, usually w pre-posttest format;