The everyday blog of Richard Bartle.
RSS feeds: v0.91; v1.0 (RDF); v2.0; Atom.
10:00pm on Friday, 10th March, 2017:
Anecdote
Today, the investigators (yes, that's what we're called) of the IGGI Doctoral Training Centre descended on Goldsmiths University of London to decide who of the 21 shortlisted applicants would form our cohort 4 in September. This means that somewhere between ten and thirteen will be happy when the email confirming their success arrives next week; the rest will be gutted.
So, IGGI stands for "Intelligent Games, Games Intelligence", and apparently it's the largest games-and-AI research group in the world. It offers full studentships and a stipend, meaning you don't pay any fees and you're given a living allowance (something like £15,500 a year, I don't know the exact figure). Because of this, it's heavily oversubscribed. The 21 candidates we interviewed today were the ones who made it through local interviewing processes at each of the three universities involved (York, Goldsmiths and Essex). As in previous years, this meant that every one of them was suitable for the programme; our job was to decide which candidates were the strongest.
The word "strongest" there suggests that it should be fairly straightforward to rank the candidates, but it's not at all. They have strengths in multiple dimensions. Some have academic publications, some have MScs or MAs, some have lecturing jobs, some have already started PhDs; some have heavy-duty programming skills, some do game jams every weekend, some have worked in the games industry; some are eloquent, some bubble with enthusiasm, some have nerves of steel, some can think fast on their feet, some are wildly industrious. Some are the finished article, some are all potential.
Of course, some have no postgraduate experience, some can't program for toffee and some are so nervous it's amazing they can say anything. They'll compensate that in other ways, of course. No-one's perfect (apart from me, obviously).
The way the successful candidates are chosen is as follows. We have three interview panels, populated with one investigator from each university. One of these will be a senior investigator and another will be an industry liaison person (I fall into that category). Each panel considers seven candidates in 40-minute sessions: 30 minutes of interview then ten minutes of discussion. 30 minutes of interview isn't as long as we'd like, but if you have to see 21 candidates in a single day, that's how it goes. All candidates are asked the same basic questions, but the depth will vary. For example, a candidate with little research experience could expect to be quizzed regarding what they think a PhD entails in more detail than someone who has an MSc, whereas a candidate with a vague or over-ambitious research proposal could expect to be challenged on the particulars of their plans.
Candidates are rated out of ten by each panel member: 10 or 9 is a shoo-in; 8 or 7 is almost certainly a yes; 6 or 5 is borderline; 4 or less is almost certainly a no. After all interviews have concluded, the members of all three panels get together and our respective marks are entered into a spreadsheet.
This is when the arguments start.
It varies each year, but perhaps five to eight candidates will average 7 marks or more. We'll take these in turn, checking that no-one thinks a particular candidate's marks aren't a true reflection of their abilities; if all goes well, we out them onto the provisional accept list. It's very rare that anyone who gets 7 or more is objected to by anyone, but it can happen.
Some four to six candidates will average below 5. We go through these seeing if anyone thinks the marks are unfair. Usually, someone will, and it may be that a candidate is moved up into the borderline zone for reconsideration. In general, though, it's not going to happen because to score this low a candidate usually has some kind of issue (say, too narrow a research focus in an area peripheral to IGGI's remit). We may encourage the candidate to reapply the following year, though, if this isn't their second application already.
We're then left with the borderline cases. We iterate through them, discussing their proposals, asking the panel members who interviewed them what they think, asking people who interviewed them locally what they think, and basically trying to establish an ordering. We don't take into account which university a candidates applied to: this is all entirely candidate-centric, not institution-centric. Sometimes, candidates interviewed by one panel may be adjudged weaker than those interviewed by another panel even if they had the higher marks; it's much less common for candidates interviewed by the same panel to switch places, but it does happen. This is the most time-consuming part of the meeting, because all the candidates have pros and cons, with the pros invariably outweighing the cons. We'll gradually move individual candidates in this group to the provisional accept list. Normally, we'll be left with three to five whom we'd like to take, but we can't take them all.
At this point, we look at candidates on the provisional accept list and decide where their funding is coming from. There are some exceptions, but in general we can take eight candidates from the UK (or EU if they've been in the UK for three years — which those who did undergraduate degrees here will have done), plus one candidate per university from the EU who hasn't been in the UK for three years, plus one candidate from outside the EU. Normally, we'll be able to find funding solutions for all the people on the provisional accept list, but it's not always possible (particularly if there are two international candidates on it); if that's the case, we have another round of discussion to make sure that the candidate we have ranked higher really is the better one, and give them preference. When it comes to the borderline candidates, we see what funding options we have left and apply them. This means that choice of university could now become a factor, because there may still be a non-UK EU grant available for one university but not for another.
In theory, this is also the one point where we can use external factors to break ties. If one of our industry advisor companies has expressed an interest in an applicant''s research proposal, we can use that as a decider. We're also allowed to take the overall gender, minority and disability balance (hmm, maybe age too?) into account here if we otherwise can't separate individuals. I don't think we've had to do that yet, but it remains an option.
After this exercise, we have a final read-through to make sure that all the interviewers accept the list of people to whom we'll be offering places. Candidates who were on the provisional accept list but not the final accept list are treated as reserves (and may well get in, because we'll usually do have someone who's offered a place but then decides not to take it up). It's possible that candidates who were borderline and then rejected could have a reserve slot, too.
Thus, life-altering futures are determined.
Acceptance letters are prepared over the next few days and the successful candidates informed. I can't tell you how many made it, because that would make their agony of waiting worse.
I can tell you that Goldsmiths do better sandwiches than we do at Essex.
Latest entries.
Archived entries.
About this blog.
Copyright © 2017 Richard Bartle (richard@mud.co.uk).