One of the most important things is just a solid familiarity with the heavily tested flaws that show up again and again. For instance:
*Drawing a cause/effect conclusion from premises that only give you correlation.
*Mixing up a necessary and a sufficient condition.
*Part to whole and Whole to part fallacies.
*Conclusions based on biased samples.
But more generally, the questions you’re always asking yourself is, “Even if all of these premises are true, why might the conclusion be false?” Very often, that involves a mismatch between the premises and the conclusion, and if you can identify that mismatch, you’ll well on your way to spotting the flaw.
What do I mean by mismatches?
*Term Shift: The premises are about *sports*, but the conclusion is about *hobbies*. Or the premises are about *cars*, but the conclusion is about *vehicles.*
*Degree of strength: The conclusion tells you that something might be true, but the conclusion tells you that it’s probably true.
*Quantifier: The conclusion tells you that something is true of *some* people, but the conclusion tells you it’s true of *most* people.
Hope this helps.