Ok guys wait a second... let me see if I can wrap my head around what's going on in this debate. and mind you I am not defending any ONE argument here but =
Monger you say Observational study > than nothing
so
that means, using your math,
that shitloads of studys, surveys, links, reviews etc. posted to support one argument > than ONE observational study > than nothing
right?
They cannot be used to indicate cause. Unless you isolate and measure/quantify the variable you wish to describe, it is nothing more than an observation.
Example: if you were going to measure insulin response vs. kcal balance to determine effectiveness of diet or % fat loss, you would need to:
1. Control the diets of all parties studied after dividing them into groups.
2. Take not only weight, but body comp tests prior to start of study, along with anything else you wanted to measure.
3. See if something is screwed up at this point.
4. Proceed, tracking everything that is consumed by both groups.
5. Proceed, measuring insulin response at pre-determined intervals.
6. Proceed, tracking weight, body comp and any other factors.
7. Track activity levels of both groups, note changes in activity levels that must be factored in.
8. Check again to see if something is screwed up.
9. Correlate data at various points within study (if it were six months, usually at every month, but this can vary depending on controls and budget)
10. Put everything together.
11. Make sure nothing is screwed up.
12. Have someone else make sure nothing is screwed up.
13. Peer review (this freakin' blows)
14. Kiss someone's ass.
15. Get it published, never once using the words fact or proof.
This can actually be more complicated, depending on who you work for and who controls the budget.