Hi Berner,

I think you can do a pre-post study: it's like an imperfect clinical trial without a contorl group. In this way, the only problem is you can't know if the effect on X depends by F or it'a a normal evolution.

I could advise to make a clinical trial with 2 randomized groups (RCT), so with an intevention and a control group (this one not exposed to F).

You should comment in the methods section your "intention to treat" (if there's a positive effect on X due to F, you can expose control group to F).

Hope I helped you.

Best regards,

Nicola.

You can use an experimental design with case control design. I wish I had more information about your variable X, and type of effect you are studying. So lets make up a situation. Say you are studying the effect of apple in boosting potassium levels in a human body. So you can design a study as a Pre-post case-control experimental design. Your cases make people who will take certain servings of bananas a part of their deite and your controls keep their dietary habits unchanged. You will need any other design only if you have a certain other fruit for comparison. Not sure why you need another another fruit with the same composition.

Thanks Nicola Sante Diciolla this was really useful!