How might you use an experimental design to test a question
How might you use an experimental design to test a question that would be relevant to your program evaluation (e.g., compare an outcome for a group that received the program versus
How might you use an experimental design to test a question
Discussion (Whole Class): Experimental Design Considerations
Experimental designs attempt to limit the possible number of alternative extraneous influences (confounds) so that what is observed as outcomes can be considered to be representative of a direct relationship between input (e.g., a training experience) and output (knowledge and skills gained through the training).
The gold standard for experiments is random assignment to condition. However, in the real world we often can’t decide who does and does not receive the training or intervention.
There are constraints on whether an intervention can be withheld from
someone (such as those assigned to a control or no treatment/intervention group) if that person might be harmed by not receiving it.
Also, ethical standards safeguard informed consent, which limits most experimental designs to those who volunteer, which then means we don’t know if the same outcomes would apply to those who would not volunteer for such a program.
By Day 3
Post your comprehensive response to each of the following:
Briefly restate the general nature of your team’s RFP’s program (e.g., school anti-bullying program).
How might you use an experimental design to test a question that would be relevant to your program evaluation (e.g., compare an outcome for a group that receive d the program versus one that did not during the same time period, especially if the participants are randomly assigne d to be in the program/no program groups)?
What is a specific question you could address for your example?
Describe the specifics of your design for addressing this question. Also, what is your independent variable?
What might be a viable dependent variable to answer the question? What hypothesis would you be testing for the proposed evaluation?
Lastly, what would be one method you might use to reduce threats to internal validity with this design?