By going through it with a fine-tooth comb.
Look at their leaps from previous research to current hypothesis
Look at their application of said research
Look at their chosen experimental paradigm and it's previous use/research
Is that paradigm suitable? What won't it show? Why is it a good/bad choice?
Look at their sampling. Representative? Bias? Big enough? How could it be better?
Go through their stats with a fine-tooth comb. Is it really showing what they say it's showing?
Possible alternative explanations? Implications on their conclusions
Extraneous/Confounding variables? Experimenter bias?
Have they clearly defined their variables?
Have they attempted to control for any errors in measurement? Did it succeed?
Have they made valid conclusions? Was it worth doing in the first place?
Has it contributed to its respective field? Why/Why not?
What questions has it raised/answered?
etc etc etc
What year are you in?