One easy way to do this is to make predictions: "If $THEORY is correct, then under the same conditions, I will observe $OUTCOME. If $OUTCOME does not obtain, I either haven't properly conducted the experiment, or $THEORY is wrong.
1. https://translate.google.com/#auto/en/assumption 2. https://en.wikipedia.org/wiki/Scientific_method
EDIT:
In response to your rephrasing of "if...then", to "assuming...then", consider that we actually do (or do not) observe $OUTCOME -- we don't assume that we observe it. To use an outrageous (but hopefully clear) example, do we know decapitation is fatal, or do we merely assume it? A long time ago (but in this galaxy) someone had a theory that decapitation was fatal. Experimentation gave us knowledge on this issue. Doubtless, they did not have The Scientific Method in mind, but they used the same process.
Let me rephrase this as:
Assuming $THEORY is correct, then under the same conditions, I will observe $OUTCOME
A theory is an assumption.
I don't want to sound like I'm attacking you here, but our beliefs about knowledge and truth are very consequential. If we cannot know anything aside from our assumptions, then all assumptions are equally valid (since we can't know which to discard). This naturally leads to situations where, e.g., racism is assumed to be good by one person, assumed to be bad by another, and there is no way we can determine who's right. In fact, "right" and "wrong" become meaningless because they're only assumptions, as does all language.