You seem to be thinking that the *entire* experiment is conducted blindly. That’s only the first part.
You perform the double-blind experiment first and then analyze the data afterward with full knowledge of what happened.
The double-blind is only for conducting the experiment and gathering the data. Once you’re done with those parts of the experiment, then you crack open the data and actually see which subject received which treatment — placebo or not-placebo. *Then* you can draw conclusions about what happened.
Basically we keep everything the same between two groups (as much as possible), and only change one variable. And we split them into groups randomly, so we don’t accidentally include some bias, like having older people in one group, or more people with diabetes.
The idea is that if you change only one thing, and all the other major characteristics are the same, then it is very likely that changes in your desired outcome are due to that one variable you changed.
There are specific statistical tests you do to see if your results are significant or not. For example you might flip a coin 100 times, and it could maybe land on heads/tails 51/49 times. This would be considered pure chance. 55/45 might also be. But 87/13 would not be. A bunch of math nerds figured out the best ways to tell if your results are likely due to chance, or maybe there is some external effect at play (like a quarter being rigged, or maybe your medicine cures the disease).
>How we know the findings are useful to apply to real life?
That’s where clinical expertise comes into play. It takes a lot of experience to read research, interpret in context of the larger body of research, find the faults and strengths, understand the actual implications of the research, and determine how to incorporate it into real life decision making. It’s hard even for doctors, and there will often be panels and debates about to best interpret the findings of particular research.
Latest Answers