How do we as researchers or as management reviewing research guard
against  correlation Fallacies?

It's not the correlation (simply the measure of a relationship) but
the implied causality.   Example: A university engineering department
wanted to predict whether freshmen would eventually graduate.   Along
with reasonable questions they included nonsense questions including:
"How do you like your carrots, raw or cooked?"   Wouldn't you know
this question provided the most significant predictor.   Answer is
raw.   Accept the correlation within its parameters and question
causality.