Although scholars spend much time and effort writing critiques of published work, and graduate students spend even more, systematic treatments of the criteria for, and strategies toward, good critiques are rare. Here are just a few points to consider.

• A critique can never be better than its author's understanding of the work being criticized.  And a reader cannot appreciate the intent of the critique beyond their grasp of the critic's reading of the original work.  This means that to write a good critical review, we must
  • first be sure we understand what we are criticizing and,
  • second, we must present a clear summary of that understanding as part of our critique.
• Although there are important exceptions, most critical reviews really are elaborations of the answer to a simple question.  After reading the piece, do I, the critic, find the argument worthwhile or not?  However elaborate or simple, with whatever style of presentation, the review is largely an effort to present a sustained defense of that evaluation.
• A good review is always fair.  We should never shy from identifying a flaw or calling a mistake what it is.  But we should always try to use the language and tone that we hope reviewers will use when they reveal the similar failings in our work.
  • We review manuscripts, articles, and books.  We do not review people.  Brilliant scholars occasionally write dim papers--they are not dim as a result.  A wise reviewer avoids referring to the author, and concentrates on the strengths and weaknesses of the work being reviewed.   For example, saying that "the arguments in the last and first sections contradict each other" is preferable to saying "in the last section, the author contradicts what she said in the first section".  Attributing thoughts and intentions to an author is even worse (e.g., "the author seems to believe", or "apparently the author is trying to ..."). Sitting in judgment of the author's intellect, effort, or morality is the worst and we should always avoid such judgments, no matter how much we may be thinking them.  Again, we review written works, we do not review people.
  • As reviewers, we commonly want to be clear about two interpretations of what has been accomplished in the work being reviewed: that of the author and that of the reviewer.  Sometimes these will be the same; often not.  A critic has no obligation to agree with the author's view of the what has been done in the work.  But the scholarly critic does have an obligation to grasp and accurately present the author's aims and orientation as they are conveyed in the work.  Once we reach the part of our review dedicated to assessing and criticizing the material, we may express an opinion that, for example, the theoretical argument or inferences stressed by the author seem a poor fit to the evidence and suggest better substitutes.  Before that, however, we should have included a summary of the author's expressed intentions and interpretation.  (Sometimes, rarely, the author's aims and arguments may be too muddled or incomplete for us to supply a concise, logical summary - in that case we can only do our best.)
  • When an author sees a review, the author should not feel (1) the reviewer has said or implied I said something that I never said or implied or (2) the reviewer has said or implied that I failed to consider something that is explicitly part of my presentation.  Good scholars do not mind critical reviews, even highly critical ones, that are accurate and fair.  Good scholars despise reviews that are inaccurate or unfair, even if they are positive.
• In general terms, most reviews include three components, although they may be merged rather than appearing separately.
  • What does the reviewed material say?  In simplest form, this is a summary.  It conveys the original author's thesis, evidence, and logic of argument.  We are telling the reader of our review what the piece achieves according to the author.
  • How does the content of the reviewed piece relate to the relevant existing literature?  With what theories or findings by others does the piece agree or complement or disagree?  We are telling the reader of our review where we believe this work fits in the intellectual or scholarly conversations about the relevant topic.  Where possible, this will show how the piece builds on other work, how it seeks to challenge other work,
  • What are the strengths and the shortcomings of the work?  Does it offer new ideas or findings, and, if so, are they compelling?  Does it address prior work in a clear, effective, and sufficiently complete way?  Does it have serious flaws of logic, evidence, analysis, or omission?  We want to remember that we should provide the evidence or logic needed to support our criticisms; we do not offer unsupported opinions.
  • Note that for the first two components - conveying what the reviewed material says and how it relates to other work and ideas - we are stressing an impartial portrait.  In the third component - the assessment - we dispense with the neutral position, developing a judgment of the quality, creativity, validity, and significance of the work.
  • Note that in the third component - the assessment - we should also include our ideas about how to rectify its flaws.  Our goal is not simply to poke holes in someone's work, but to offer ideas (when we have them) how we can make the best of it.  We can also discuss how others might build on the work being reviewed, but this is more the original author's terrain, so we need not feel obligated to pursue this.
  • It may seem daunting to review a scholarly work without authoritative scholarly knowledge in the relevant fields.  Unavoidably, the depth and effectiveness of all three components, but particularly the critical assessment, will depend on the reviewer's authoritative knowledge and confidence.  Even when we lack authoritative knowledge in a field, however, so long as we could reasonably be considered as part of the intended audience, we can write thoughtful, accurate, and useful reviews (we can also sometimes do the job when we are not the intended audience but we will skip that here).  If we are part of the intended audience, then any aspect of the work that is not presented in a way we can follow and understand must either be superfluous to our assessment (i.e., the author may include material intended only for "experts" and their special concerns, with the understanding that the main argument does not depend on understanding this material) or reflects an error in presentation by the author (e.g., if the author presents a theoretical or methodological component section that is necessary to the overall argument but left incomprehensible to the a large part of the presumed audience).   This means that we can responsibly and effectively review the work, so long as we explicitly "bracket" the aspects of the original material that we will neglect.
• Our critiques can be stronger and more useful in general if we consistently distinguish between several levels of error and their implications.
  
  • Errors of presentation.  It is commonplace for authors to make mistakes.  They leave gaps in arguments, misconstrue the meanings of other authors, neglect an implication of the data analysis, and the like.  These mistakes have little significance to the argument; they are simply embarrassments.  These are mistakes that we expect the author would agree about and correct if given the chance.   All of these errors deserve to be pointed out, but we want to recognize that most of them do not challenge the validity or value of the argument that has been made.  Most errors of presentation are just that.  They detract from the polish or the completeness of the work, but do not diminish its plausibility or merit.
  • Significant but correctable errors.  More serious flaws appear when the author makes a mistake of logic, evidence, or omission that weaken the the main argument or theory, or even, in the worst cases, prevent it from working or making sense.  The question facing the critic is this: is it possible to correct the error without changing the identity of the theory or argument?  The first challenge for the critic is to show that the author's argument is seriously hampered by this error: its predictions are unreliable, its claims contradict knowable facts, it contains internal contradictions or ambiguities that prohibit rigorous application, or the like.  This is often a difficult task (after all, published works commonly reflect long and hard work on the arguments, rely on considerable authoritative knowledge, and have survived examination by other authoritative reviews).  But the harder task - which should be compulsory at least to attempt but commonly is not - is to suggest a repair to the original argument that would overcome the shortcoming. Showing how to repair a significant error is a scholarly contribution beyond the common goals of a review, and should be highlighted as such.
  • Fatal flaws.  A critical flaw in a theory, argument, or claim is an error that cannot be corrected without changing the nature or identity of the original argument.  If the original argument relies on an empirical claim that is wrong, or if the inherent logic of the argument is internally inconsistent or otherwise unworkable, then it is fatally flawed.  Only a fundamentally different theory, argument, or claim has a chance of validity.  Critics often seek to unearth fatal flaws but rarely succeed.  Critics often present correctable errors as fatal flaws, sometimes because the critic cannot see the strategy for fixing the error any better than the original author; sometimes just because the critic hopes for the status gain associated with finding a fatal flaw.  We always want to avoid making a serious error while we are arguing that someone else made a serious error.
  • If we systematically consider at what level each flaw lies and adopt a fitting response, our assessment becomes more accurate and more accessible.  It is surprisingly easy to confuse ourselves, and once we do that we are certain to confuse others.