Answer:
A.
Ok, suppose an experiment where you have a lot of dependent variables.
something like the variables y, z and k that change when you change the variable x.
Now, as you play with the variable x, you will notice changes in the variables z, y and k.
But there is now a question (a lot actually)
if the change in k was actually caused by the change in z?
Then you need to try to find a way of changing the value of z for a fixed value of x (designing an experiment) and see how k reacts.
This may be really hard, so as few responding variables you have, the simpler is the experiment, and easier will be to collect good data.
Now we also could discard the other options:
B: If the procedure is complex, then there are more ways in which the scientist can make a mistake that will introduce error to the data, so this will not improve the quality.
C: A larger sample will allow you to have more precision, as more subjects are being observed.
D: When we manipulate one variable, we may be actually introducing some quantity of error to the data.
For example, if the quantity is the voltage for an experiment, we actually do not know the exact voltage tath comes from the source, so there is a small source of error. Now as we start adding more and more independent variables, we also are adding more of these sources of errors, so an experiment with several independent variables will not improve the quality of the results.
