Here is an approach.
- Skim. First skim over the reading to get a sense of
the themes it covers.
- In one minute, a good abstract should give you a sense of the goals
of the work. Before reading further, jot down what questions you hope
the reading will be able to answer for you.
- Next, read the introduction and conclusion. This is normally enough
to get a sense of the big picture. By now you should know what the main
claims are going to be. Ask yourself: Are the claims in the text
surprising? Do you believe them? Can you think of examples of cases that
do not seem consistent with the logic of the argument? Is the reading
answering the questions you hoped it would answer? If not, is it
answering more or less interesting questions than you had thought
of?
- Predict: Ask yourself: What types of evidence or
arguments would you need to see in order to be convinced of the results?
For an empirical paper try to think of the sort of data you might want
to get. For a formal paper try to sketch out the contours of a
model.
- Read the heart: Only now read through the whole
text, checking as you go through how the arguments used support the
claims of the author. Of course you hardly ever really read the whole
text. Instead:
- For many empirical pieces you read by going first to the tables and
figures and assessing the claims and evidence, then check intro or
conclusions for interpretation and other issues. Then check any parts of
the strategy that seem unclear. Do not rely on the authors
interpretation of tables.
- For theoretical pieces you read by going straight to the
propositions and theorems.
- Skip the literature. I read the literature review last if at all.
You want to check that the article is making the contributions it says
it is, but the literature review is sometimes only of limited help for
this.
- Re-create: In all cases when it is possible see if
you can download data, replicate results and use it to probe and test
the arguments. For theoretical papers try to simulate results to get a
feel for the moving parts.
- Assess: It is rare to find a piece of writing that
you agree with entirely. So:
- As you come across issues that you are not convinced by, write them
down as points for discussion.
- Also note when you are pleasantly surprised, when the author
produced a convincing argument that you had not thought of.
- For a checklist of things to look out for as you read see here: checklist