Review is most helpful when you understand the general, structural point of a question - not just an abstract understanding.

Did you identify the question type correctly? If not – or if you’re in the habit of just generally trying to answer the question posed without classifying it – identify the specific words in the prompt that should have alerted you, and write them down to review later. 

Did you pick up on the important words/phrases in the stimulus? You should be looking for different things depending upon the question type, but here are a few examples: If it’s a Must Be True question, then comparative language is important – “better than,” “smaller than,” etc. If it’s a Flaw question, then there should be language related to a prevalent flaw. When you figure out what that language is, write it down to review later. 

How could you have anticipated the right answer if you’d done the above steps correctly? It’s important for both speed and accuracy that you have an idea of what the right answer will look like. Anticipation also depends on question type. For instance, if you have a Must Be True question with comparative language, e.g., “Bob is worse at tennis than Jim,” then you can anticipate that the answer will flow from that comparison, e.g., “Bob is not the world’s best tennis player.”

Finally, why was the wrong answer wrong? Just like the second and third steps, this depends on the question type. For each question type, there are criteria the right answer must meet, as well as common ways answers don’t meet those criteria. For example, the fact that an answer falls outside the scope of the stimulus – refers to something not stated in, or inferable from, the stimulus – is disqualifying for a Must Be True answer, and many wrong answers are wrong on that account.