Is it common to review papers assigned to your supervisor?

It is not uncommon for the advisor to ask a student to review a paper. Let's face it — reviewing is a chore, takes a huge chunk of your time and most people do it not out of interest for the paper, but rather a necessity (it factors in promotions within the organization and recognition within the society) and probably some feigned interest in the "role of an academic" and "giving back to the society".

Usually, the papers that an advisor passes on to their students are

• tangentially related to their interests
• only mildly interesting or borderline stuff
• are not from top-tier journals

As a student, you should never pass up on these opportunities. It can be a very rewarding experience to review a paper and discuss with your advisor. Some of the things you'll learn are:

• that you didn't catch on to some inconsistency/mistake after a month of review, but your advisor did it in 15 mins of casual reading.
• to write an effective critique while at the same time not putting the authors down. Contrary to what many may think, a review is not all about "your math sucks!". While yes, it should be pointed out if it indeed sucks, there are ways of saying that. If you spend a lot of the review finding faults, then it is also "fair" to throw in a couple of nice things about the paper and balance it.
• you'll also learn to not nit pick on minor quibbles and focus on the technical content (I hate it when a reviewer makes a huge deal of 'data are' vs 'data is').
• You'll learn to do this while juggling other activities, so it's also an early lesson in time management.

In the end, the quality of the review is partly your advisor's responsibility and he's not going to allow it if it is completely shabby. So, take this as an opportunity to learn about how to write a review.

Speaking for myself, I reviewed a couple of papers for my advisor in my early years and now, after I have a few publications to my name, I get invited to review papers on my own (i.e., not via my advisor). Recently, there were a couple of opportunities that my advisor passed on to me, but the difference this time, is that he trusts me to do a decent review and so requested the editor to officially ask me to review, so that I'll get credit for it.

For a conference, PC members are responsible for reviewing papers. However, it's very common that they can use sub-reviewers, that is, delegating their review to another person. They nonetheless remain responsible for the final review. Some reviewing system, such as Easychair, have an integrated mechanism for asking a sub-reviewer, so there is an official "proof" of the review delegation, and some conferences include in their proceedings the additional reviewers (without saying for which paper, of course).

As for the frequency of such practice, it's hard to tell. Some supervisors might easily use that as an easy way to not do the reviews, others can use them because they believe the student might be more qualified (and more interested) in the particular topic of the paper, others can do it because reviewing is an important task of an academic, and one has to learn at some point. However, if the student does not feel qualified enough to review the paper, he/she should mention it and refuse the review.

This is partly a response to john k's answer.

In the end, the review will go out only under your advisors [sic] name

Farming out a review to someone else -- especially someone who is subordinate to you -- and then putting only your own name on it is a practice that the vast majority of academics would view as unequivocally unethical. In fact it is essentially plagiarism.

I wish answerers here would not endorse this practice. As Jeff E points out in a comment, if this is done only between the student and the adviser then the editor will not even know about it so certainly cannot be viewed as allowing or participating in this.

So, take this as an opportunity to learn and do not worry that you won't get credit for it.

I certainly do worry about the practice of advisers passing off their students' work as their own.

and he's [the advisor] not going to allow it if it is completely shabby.

If the advisor is going to be involved in the refereeing process anyway, then passing it off to the student is not a clear time savings. However it sounds like a recipe for a bad refereeing job: one very junior, possibly not fully qualified person actually reads the paper, and then the more senior, qualified person mostly looks at the report that the junior person has written. There's a lot of room for something to fall through the cracks.

Here is what I would suggest instead:

1) If as an adviser you actually feel that your student is fully qualified to referee a paper that you have been sent, and if you feel that it will in some way be more useful for your student to do it than you yourself (e.g. you view it as a learning experience for the student), then contact the editor and suggest your student as an alternate referee, while identifying them as your student. Most editors will be happy with this because:

(i) Graduate students can make unusually conscientious and quick referees. I refereed one paper as a graduate student. I was very flattered to be asked to do so, so I carefully read every line and even spent a substantial amount of time gaining background knowledge [for the mathematicians: I had to read part of SGAVII in order to do what I felt was a thorough job]. I also did this over a period of slightly less than a month.

(ii) It's better than asking a very junior person to do the refereeing job out of the blue [as in fact was the case with me in the above situation] because the student can consult with her adviser if necessary, either to ask a key technical question or just to ask for advice about writing the report and dealing with the editor.

I (formally) passed off a referee job to a PhD student of mine about a year or so ago. I think it worked out well: he did a good, fast job, and he did consult me a bit for help. As I recall, most (or all) of the questions he asked me were not about the content of the paper -- on the contrary, after he read it I asked him some stuff out of my own interest and curiosity -- but rather about the mechanics of the referee report: what standards to impose, how to reply to the editor, whether it was okay to request that the authors make a revision doing X and Y, and so forth. I think this latter part was at least as valuable an experience for him as simply vetting the paper for correctness.

2) If you can identify a part of a paper that is sufficiently independent of the rest so as to make it possible to read and judge only that part, then it seems reasonable to contract a "subreviewer" for that part. Maybe this subreviewer is your student and asking her to do it is saving you some tedium. Maybe this subreviewer has expertise that you lack and it is because of them that you feel capable of writing a report at all. With respect to this, I would say:

(i) Of course you need to identify by name all subreviewers that you use.

(ii) If you take on a subreviewer, it should be very clear exactly what the subreviewer is vouching for. You are still the one who is vouching for a paper as a whole. You need to be extra careful that you are not leaving any gaps. For instance, if a paper concerns arithmetic geometry and spectral graph theory, then there will probably be a part in which problems in arithmetic geometry are translated to problems in spectral graph theory, or vice versa. There needs to be at least one referee who understands both of these topics sufficiently well in order to do a credible job.

(iii) If the issues in (ii) are at all complicated -- e.g. if you find yourself wanting to enlist more than one subreviewer -- then it may be better to have this subcontracting process done by the editor. S/he can then, if desired, solicit multiple reports, add/change/remove subreviewers, and so forth.

I have never enlisted subreviewers, and I have refereed more than 30 papers. But I could imagine for instance being given a paper which had a substantial computational component, and then I might ask to see the code itself and give it to a student of mine who has substantial programming expertise. The difference here is that doing so would improve the referee process (it is very unlikely that I would look at the code myself in any detail, and even less likely that any good would come of my doing so). Passing something off to graduate student when you think or know you would do a materially better job already seems slightly ethically suspect.