0
$\begingroup$

My research question is: ''Does alcohol usage and breathing exercise's effect on reducing excitement differ in women and men ?''

I have 2 within-subject variables: breathing and alcohol use, and 1 between-subject variable, gender. I want to do pre/post experiment which has high internal validity. I think that I can use pre test- post test control group design but I couldn't use this design. How can I use control group with these within subject design ? If I use control group, will I distort the within subject variables in my design ?

$\endgroup$
  • $\begingroup$ Validity is an aspect of a test. If you want to assess validity, you would have to throw out post test results because they are influenced by the intervention in the non-controls, or by learning effect/fatigue in the controls and non-controls alike. $\endgroup$ – AdamO Dec 13 '18 at 23:01
0
$\begingroup$

The validity of your measure of "excitement" is relevant but you cannot assess it with the trial data. Doing so compromises the study. That's because the study protocol has said to analyze excitement as if excitement were measured by a valid instrument: I only know this because if your excitement scale were not valid, the study design would be invalid. Analyze the trial data using the proper ANCOVA design for pre-post with control.

Refer to the existing research if you must comment on the validity of the scale. Pay particular attention to differential item functioning by sex. If no research exists, conduct a separate, ancillary study to determine validity of your excitement scale. Do not recruit subjects from the initial study. The considerations are completely different from sampling strategy, to incentive bonus, to power calculation: it is a completely separate objective.

You cannot assess instrument validity post-hoc.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.