What do we look for in an applicant to the lab?
I was asked this question recently in a podcast done by the institute. This is a podcast that replaces "reflection talks" that faculty used to give before their 5y contract renewals or promotions.
I used to accept people to the lab simply on the basis of an interview to gauge interest. I thought the rest could be discerned from their CV. Unfortunately, this quickly backfired because CVs don't reflect reality. More often than not, CVs are exaggerated. In other cases, they are filled with internship experiences but each lasting a month or two. In other words, there is a distinct lack of depth and commitment shown in most CVs but people are quite articulate these days and a simple interview is insufficient to gauge preparedness to take on research in cognitive science.
The other issue is people think research in cognitive science is easier than say in quantum physics. So many people who have zero exposure to the field apply because they believe that cogsci research is essentially philosophizing or thinking about thinking. Which ought to be fun and easy since many of us do engage in such thoughts from time to time. When they enter our labs, they get a dose of reality. Research is a hands-on and mundane endeavor like almost any other pursuit. There is a method to the madness and you must be willing to commit to doing the small and boring stuff day in and day out for several years before you see tangible rewards in terms of publications or fellowships. So unless you have some sense of what it takes, you are likely to be disappointed once you start this journey. Unless you enjoy this difficult and boring process of the pursuit of knowledge and pushing the boundaries of what we know with serious intellectual and hands-on struggle, it is not possible to make much progress and you'd be wasting your time pursuing a research career.
Now the most prestigious grad schools in the US and other places accept candidates who have already published papers in top venues. So they have some strong indications that these candidates understand what to expect. The situation in India is very different. However Indian labs have a very important role to play. People who haven't managed to publish by the time they apply to a PhD program aren't necessarily any less talented than the people who manage to go to the best universities abroad. They've just not had the access to high quality mentorship and resources to get to that point.
So in this context, if publications or other indications cannot (and should not) be used as a signal of preparedness, what can we do? I have heard some senior faculty say "the entry bar must be low and the exit bar must be high" but I completely disagree with this approach. I understand that the intention is to give everyone a chance but not everyone is served well by steering them to research programs that would only lead to lost time, opportunity costs, and frustrations allround. That said, I acknowledge that such an approach might work for some labs that are well established, have produced a body of work that is respected, and can afford to publish lower quality work to help someone graduate who was unprepared or unmotivated to do a high quality research thesis. This advice is disastrous for newer faculty trying to build a respectable research program, in my opinion.
So the middle-ground solution I've come up with is to give people a dataset, some sample research questions that can be answered with the dataset, a textbook with specific chapters indicated that can help them with the analysis and other resources. They take a few weeks on this problem. Then we meet to discuss. The goal is not to test whether they can find the answer or solve the problem. The interview looks into how they think about a problem. Are they curious and careful enough to question assumptions? Do they themselves make unwarranted assumptions such as assuming a column in the dataset stands for say start time of the memory probe in an experiment without digging into various log files to confirm? I don't care about AI assistance for the analysis but this is the type of question I ask them. I might also ask them how they tried to understand an analysis they did with AI assistance. Why this set of parameters for that function call? If someone displays integrity and openly acknowledges what was beyond their ability to understand, while at the same time displays some signatures of curiosity (both about the larger cognitive questions and about the analysis methods) and carefulness, that's usually sufficient for me to issue a positive decision.
This single step added in the interview process has led to the lab having motivated students and researchers who understand that we have a high bar for rigor and ambition with our research projects. It has solved all of the problems we faced in the initial few years with various personnel issues, people (esp in the dual degree program) going away without having done any reasonable amount of work towards their thesis, etc. We've had a 100% graduation rate since, with most of them being on time. Only when the DD students themselves agree to work on more ambitious journal submissions have we had delayed graduation but they too have very good outcomes due to having higher quality research outputs on their CVs.
The utility of this approach is demonstrated by the observation that DD students motivated by an expectation that cogsci research was "easier" used to approach our lab a lot. No amount of me verbally telling them that it wasn't true would change their minds but 8/10 applicants disappeared as soon as I started issuing this data analysis challenge. Since then, the number of applicants has been lower but almost everyone engages in the data analysis challenge. This is some indirect evidence to me that word has gotten around that one needs to be genuinely interested and willing to do the boring and hard work day in and day out to succeed in this lab.
I have also seen huge red flags in PhD applicants thanks to the data analysis round where they clearly answer my live questions using AI assistance but cannot answer very basic Qs that can be answered only if they've looked into the data and methods themselves. Integrity is number one. They can have multiple publications but if I'm not confident in the integrity they display, it's going to be a hard no.
Hope this helps clarify the process for future applicants. I'm also hoping that this helps new faculty at IIITH to decide how to set up a robust process for assessing applicants to their labs.
Comments
Post a Comment