Thursday, January 15, 2015

Choosing a research advisor

How can you avoid a picking a bad adviser if you’re not sure what you’re looking for in the first place?
One way is to study what goes into the decision. Be an informed consumer: Know what you want, and expect what you’re entitled to.
From the comments
Spotting (and avoiding) a bad adviser is actually much easier than finding a good one. Common issues to watch for:
(1) Ask: find former students or colleagues and [delicately] ask whether (s)he is an effective adviser who supports students. 
(2) Speak: is (s)he someone with whom you can truly speak... candidly? You will need that later on, so avoid those aloof, cagey, dismissive, and distant personality types (and yes I realize that probably rules out half of academia);
(3) Publications: does (s)he publish results regularly? If not, avoid. Remember you're there to do work and get published. 


Anonymous said...

In IIXs, sometimes students choose advisors who will not ask them to work and they can enjoy a good life for six years. after that they realize, it is not so rosy after all.

Anonymous said...

The three points that you have mentioned may not work while choosing an adviser in USA.

1) Asking former students or colleagues may not work because most of the students, in trying to be professional, avoid speaking ill of anyone, especially their advisers. All you can get is some set of vague statements from which it is hard to extract information.

2) Most of the professors are well mannered in class. You can distinguish between them only after you have personally worked with them and fail to do good work. The true nature of an individual often comes out when they are unhappy. But, there is no way you can get these kind of information before choosing an adviser, especially when point (1) does not work.

3) Every serious researcher has to publish in order to survive and justify grants, and also to get promoted. Thus, almost all the professors now a days publish. I think one really needs to judge the quality and/or nature of work (theoretical vs systems) to choose a professor.

These are based on my personal experience, as an ex-student of IISc and also after having done my PhD from USA.

Anonymous said...

The points mentioned and the article are for choosing advisors in USA but also India. But the article and comments are basically for USA.

#1. If the students avoid answering you or do not speak greatly of the advisors, then you can interpret that it is not positive.

From the comments,

1. Do they keep predictable office hours? 2. Do they return email in 48 hours? 3. Have they published a refereed manuscript in the past 3 years? 4. Do they have at least two other advisees? 5. Is their work well-cited? 6. Have any of their advisees become successful? If the answer to any of these questions is 'no" -- proceed at your own considerable risk.

The above is more applicable in India.

Senior IISc Prof.

Anonymous said...

I am the person who has posted the comment at February 2, 2015 at 4:19 AM. I would like to comment on the post of Senior IISc Prof. Sir, thank you for your comments.

But, I have not met any student who has ever spoken greatly of their advisers. Somehow being professional includes not speaking openly about their advisers, good or bad. It is perhaps because an individual may not want to seriously influence a young person's decision making. Imagine if I say great things about my adviser, encouraging him to join, and then he realizes they are not compatible. I personally would not want that, so I would choose to give a more balanced response. All the people I have talked to do the same. I do not quite agree with all the other points in the post as well.

During my PhD, and after interacting with other PhD students at various universities in USA, I feel that the concept of a good adviser does exists but it is highly subjective, depending on a person's own style of working and thinking.

Let me be more specific. I know one professor who mostly likes to behave as a manager for their students, wait for them to think deep and take the initiative in problem formulation, and then guide them by telling them if what they are thinking is right or unlikely to lead to a solution.

Another adviser, my friend's adviser really, he would sit with his students for hours, teaching them how to think, sometimes spoon feeding them encouraging them. Finally leaving the students on their own in the final years. So really closer to leader than a manager.

Some advisers would not like to meet unless you make progress in your research. Others will always meet and just chit-chat to keep the thoughts flowing.

Some advisers will scold you if you are not making much progress. Others may politely cut your funding.

Also various people have various philosophies: whether they want to ignore rigor for the sake of engineering, how they look at other people's research. Some guys respect all kinds of research, some are very outspoken about what kind of research should be done.

Some professors slow down in terms of learning new tools and innovate based on what they know, some keep learning new things.

I am just putting out a lot of thoughts out there. And I am just trying to say is that perhaps these are the factors that we should really look at. The points highlighted by others look a bit over simplified.

So, in summary I am saying that to choose a good (read compatible) adviser, one needs to know their own style of work first.

One final point. Now when I look back at my PhD, I think I really wanted an adviser who is really rigorous in his approach to research, encourages his students to aim for deep math based ideas, and from whom I could have learned doing deeper mathematical research. When I joined my PhD my adviser was too rigorous for me, based on my background. But, after having worked with him, I realized he is not rigorous at all, mostly treating rigor as abstract nonsense. I don't think most of the help books helping people to choose adviser talk about such subtle points (assuming the community treat these as subtle).

Anonymous said...

thats absolutely right ,because its more important that what is the type [s]]he work than reputation...the style work and experience are also great factors....

Anonymous said...

Top most reason should be your research interest. You are going to make your career on the basis of this training. Most of the comments are advising people to chose some successful scientist, in other words if I am interested in working on Birds but the best people available in my options work on fishes, I should go to work on fishes.

Probably you will publish great on fishes and even get a job to work on fishes but birds will fly away from you.

All other advises are good to select the best person amongst your research interests but if you want to be a leader in research follow your research dream not your PIs research dream.

Abhishek said...

Properties of Good research advisor is valid across continent.

Look for generic features:-

1) Research topic
2) Publish regularly -
a) count number of papers of his previous students - with him
b) If a "full professor" is taking first authorship in several publications, you must be careful of this person.
c) Working with international collaboration is also important

d) Publication record below 5 per year is risky sign with more than 15 co-workers in the group.

e) Publishing in Nature and Science is dream of all researchers however, Not all can get it, hence if young researchers not having publications in 3-5 impact factor journals will not get any chance to work later. Hence, a good mentor will publish their works in such journals rather then aiming all the time >10 impact.

3) Check where his/her former group members are - if more 50% are in good academic jobs, you can consider a good sign or else avoid this choice
4) Obviously working with highly citable or reputed groups helps you in your career building.

5) Also important to understand your profile - how it will fit into the requirement of the research work you are going to do.

6) Also look grant history of the futuristic supervisor.