Dear Students
We discussed about design consideration and focused on Pre-Test Post-Test control group design. What are the problems with this design ? Please mention case where you would not like to use for your studies.
Also, there will be great appreciation in terms of grade for those who can initiate a new thread!
Pre-Test Post-Test control group design, although, improves internal validity, as discussed in class, it sacrifices external validity. It is hard to properly understand the influence of pretest on the untreated/placebo group due to a lack of prior measurement available.
ReplyDeleteFurthermore, the ethical concerns are important. One cannot isolate a group completely, especially in longer studies. Selection bias may also occur.
Social validity threats of "compensatory rivalry" and/ "demoralization" might occur in the untreated group, affecting the final result of the treatment.
Understanding and keeping in view the given problems, pretest - post-test control group design can be a very useful tool.
This comment has been removed by the author.
ReplyDeleteFollowing are some of the problems with this otherwise useful design,
ReplyDelete1. Ethical concerns
As we talked about in class, there can be ethical concerns while putting some people on treatment, and other on placebo.
2. I think demoralization is also an important aspect, where people will simply lose hope and perform worse than they would have otherwise even without treatment.
3. External validity issues of reactive interaction (bidder example), selection biases, and multiple treatment interference can also be a cause of concern.
1. The main problem with this design is that it improves internal validity but sacrifices external validity. There is no way of judging whether the process of pre-testing actually influenced the results because there is no baseline measurement against groups that remained completely untreated. As discussed in class with an example of treatment and placebo.
ReplyDelete2. The other major problem is selection biases which usually arise in some sociological and educational research programs.
For example: If the children are drawn from different schools,the chance of selection bias arises, because randomization is not possible in this case.
The problems of classical randomized comparison group design(pretest-protest control group)are:
ReplyDelete-testing Bias
-threats to validity
-behavioral changes of experimental groups