Question

When there are parts of an experiment that scientists cannot carefully control, they should

Answer 1

keep trying
adjust other vairables
or just try to control it.

Answer 2

Yes. In general, they should keep all conditions which could possibly affect the result of an experiment the same - except for the one they are trying to test (as in differing concentrations of a chemical, for example). That way they know that any change in the result is due only to the test variable and not to so called 'covariates' beyond the scope of the experiment.