In our field of orthodontics, we use the same subjects as control and experimental groups in the following way, pretreatment levels are taken as patrol and then once the treatment is rendered, as second time point , they are taken as experimental group. You may check if it applies to your field

the same test used in a different context (different operator, different subjects, different circumstances, at a different time) may not be valid. In other words, validity evidence applies to the data generated by an instrument, not the instrument itself. Validity can be equated with accuracy, and reliability with precision

Using same subject and control improves matching.

However can have bias if the learning u r studying has long term memory

Yes it is possible in crossover study.

The participants in both groups can act as controll as well as experimental subjects.

Mam in our field of orthodontics, we have split mouth design studies in which both control n experimental subjects are the same. Plz add a bit more info pertaining to your study so that we can have a better deliberation.

Ask yourself the question, will the "experimental" or "control" intervention have any impact on how the candidate will behave when they are transferred to other group? In those cases, you may have to use "cross-over" type with changed order of exposure to see if the "order" of exposure to your teaching has any impact.

In dentistry a lot of "split mouth" and cross over trials are done... The same validity may not be relevant to education because the process of education does not only impart what is taught. The student also learns other things which can impact future education... For instance, for a person with a good musical ear or training, the process of learning an entirely new musical instrument would be easier than for someone who has never learnt the same. A surgical resident who has experience playing with game controllers find it easy to navigate robots than someone who has no such experience... but eventually all reach the same level of competence. (if we observe them long enough which we mostly won't!)

Another way will be to do a Study within a trial (SWAT) or Trial within a trial approach. This is recommended in surgical cases where the different combination of exposures will constitute a group which you can compare to get inferences.

May need careful analysis and exploring "independence" or inter-dependence of factors and use of random effects model in regression. But, during interpretation, you will have to reconsider all the other possible explanations before suggesting your explanation is valid.

In short, it IS possible to do a valid single group intervention trial and it is also very much possible to do an invalid randomized trial which is methodologically sound but violates the real life scenarios. In that case, the ecological or external validity will be severely affected.

I am interested in complex or 'wicked problems' in research methods and would be open to discussion on this with anyone interested. Please send a message across as you please.

Regards,

Akilesh. R

Using the same subject as control and experimental group will depend on the study design. If it is a crossover study, then it may be fine as different groups would undergo different interventions or control at various time points. However, if you have only one group and use it as control and experimental group, then it may lead to issues regarding the confounding factors. The groups may not be of the same age when treated as control and experimental. Furthermore, growth may play a role in giving you different outcomes which may not necessarily be due to treatment or experiment. The upside is that the control and experimental groups would be matched perfectly as they are the same group.

I have the same issue with this type of experiment. I agree with crossover study perspective. Another important think we need to do is in reporting this experiment. We need to state clearly the limitation and the weakness in our experiment, and the reason why should we use this type of experiment. So, the other scientist can improve the experiment.