Threshold concepts, learning and Pokémon

Last academic year I took a course on teaching and learning in higher education. I enjoyed learning some education theory: I could recognise habits (both good and bad) my students and I practised. I wanted to write up some of the more interesting ideas I came across, I’ve been kept busy by other things (such as writing up the assessment for the course), but here’s the first.

Pokémon PhD

My collection of qualifications.

Threshold concepts

Have you ever had that moment when something just clicked? Perhaps you’ve been struggling with a particular topic for a while, then suddenly you understand, you have that eureka moment, and you get a new view on everything. That’s one of the best moments in studying.

Threshold concepts are a particular class of these troublesome concepts that have a big impact on your development. It’s not just that these take work to come to grips with, but that you can’t master a subject until you’ve figured them out. As a teacher, they’re something to watch out for, as these are the areas where students’ progress can be held up and they need extra support.

Being a student is much like being a Pokémon. When you start out, there’s not much you can do. Then you practise and gain experience. This can be difficult, but you level up. (Sadly, as a student you don’t the nice little jingle when you do). After levelling up, things don’t seem so hard, so you can tackle more difficult battles. Every so often you’ll learn a new technique, a new move (hopefully you won’t forget an old one), and now you are even more awesome.

That’s all pretty straightforward. If you keep training, you get stronger. (It does turn out that studying helps you learn).

Mastering a threshold concept is more like evolving. You get a sudden boost to your abilities, and now you can learn moves that you couldn’t before, perhaps you’ve changed type too. Evolving isn’t straightforward. Sometimes all you need to do is keep working and level up; other times you’ll need a particular item, to learn a special move, to hone one particular aspect, or be in the right place at the right time. Some people might assimilate a threshold concept like any other new idea, while others will have to put in extra time and effort. In any case, the end effect is transformative. Congratulations, your Physics Student has evolved into a Physicist!

Do di do dum-di-dum-di-dum!

Educational evolution. Pokémon art by Ken Sugimori.


Every discipline has its own threshold concepts. For example, in Pokémon training there’s the idea that different types of Pokémon are have advantages over others (water is super effective against fire, which is super effective against grass, etc.), so you should pick your Pokémon (and their moves) appropriately. Threshold concepts share certain attributes, they are:

  • Transformative: Once understood they change how you view the subject (or life in general). Understanding Pokémon types changes how you view battles, if you’re going to go up against a gym leader called Lt. Surge, you know to pack some Ground types as they’re good against Electric types. It also now makes sense how Iron Man (obviously a Steel type), can take on Thor (an Electric type) in The Avengers, but gets trashed by some random henchpeople with heat powers (Fire types) in Iron Man 3.
  • Irreversible: Once learnt there’s no changing back. You know you’re going to have a bad time if you’ve only packed Fire types to go explore an underwater cave.
  • Integrative: Having conquered a threshold concept, you can spot connections to other ideas and progress to develop new skills. Once you’ve realised that your beloved Blastoise has a weakness to Electric types, you might consider teaching it Earthquake as a counter. You’ve moved on from just considering the types of Pokémon, to considering their move-sets too. Or you could make sure your team has Ground type, so you can switch out your Blastoise. Now you’re considering the entire composition of your team.
  • Troublesome: Threshold concepts are difficult. They may be conceptually challenging (how do you remember 18 types vs 18 types?), counter-intuitive (why don’t Ghost moves affect Normal types?), or be resisted as they force you to re-evaluate your (deep held) opinions (maybe Gyarados isn’t the best, despite looking ferocious, because it has a double weakness to Electric types, and perhaps using your favourite Snorlax in all situations is a bad idea, regardless of how huggable he is).

Using these criteria, you might be able to think of some threshold concepts in other areas, and possibly see why people have problems with them. For example, it might now make more sense why some people have problems accepting global warming is caused by humans. This is certainly a transformative idea, as it makes you reconsider your actions and those of society, as well as the prospects for future generations, and it is certainly troublesome, as one has to accept that the world can change, that our current lifestyle (and perhaps certain economic activities) is not sustainable, and that we are guilty of damaging our only home. The irreversible nature of threshold concepts might also make people resist coming to terms with them, as they prefer their current state of comfortable innocence.

Loss of Arctic ice over 15 years

National Geographic atlases from 1999 to 2014, showing how Arctic ice has melted. At this rate, ice type Pokémon will be extinct in the wild by the end of the century (they’re already the rarest type). It’s super depressing…


Threshold concepts are key but troublesome concepts within a discipline. If you want to be the very best, you have to master them all. They are so called as they can be thought of as doorways, through which a student must step in order to progress. After moving passed the threshold, they enter a new (larger) room, the next stage in their development. From here, they can continue to the next threshold. Looking back, they also get a new perspective on what they have learnt; they can now see new ways of connecting together old ideas. Students might be hesitant to step through because they are nervous about leaving their current state behind. They might also have problems just because the door is difficult to open. If you are planning teaching, you should consider what threshold concepts you’ll cover, and then how to build your lessons around threshold concepts so no-one gets left behind.

I especially like the idea of threshold concepts, as it shows learning to be made up of a journey through different stages of understanding, rather than building a pile of knowledge. (Education should be more about understanding how to figure out the right answer than knowing what it is). If you’d like to learn more about threshold concepts, I’d recommend browsing the resources compiled by Michael Flanagan of UCL.


BritGrav 15

April was a busy month. Amongst other adventures, I organised the 15th British Gravity (BritGrav) Meeting. This is a conference for everyone involved with research connected to gravitation. I was involved in organising last year’s meeting in Cambridge, and since there were very few fatalities, it was decided that I could be trusted to organise it again. Overall, I think it actually went rather well.

Before I go on to review the details of the meeting, I must thank everyone who helped put things together. Huge thanks to my organisational team who helped with every aspect of the organisation. They did wonderfully, even if Hannah seems to have developed a slight sign-making addiction. Thanks go to Classical & Quantum Gravity and the IOP Gravitational Physics Group for sponsoring the event, and to the College of  Engineering & Physical Sciences’ marketing team for advertising. Finally, thanks to everyone who came along!


BritGrav is a broad meeting. It turns out there’s rather a lot of research connected to gravity! This has both good and bad aspects. On the plus side, you can make connections with people you wouldn’t normally run across and find out about new areas you wouldn’t hear about at a specialist meeting. On the negative side, there can some talks which go straight-over your head (no matter how fast your reaction are). The 10-minute talk format helps a little here. There’s not enough time to delve into details (which only specialists would appreciate) so speakers should stick to giving an overview that is generally accessible. Even in the event that you do get completely lost, it’s only a few minutes until the next talk, so it’s not too painful. The 10-minute time slot also helps us to fit in a large number of talks, to cover all the relevant areas of research.

Open quantum gravitational systems

Slide from Teodora Oniga’s BritGrav 15 talk on gauge invariant quantum gravitational decoherence. There are not enough cats featured in slides on gravitational physics.

I’ve collected together tweets and links from the science talks: it was a busy two days! We started with Chris Collins talking about testing the inverse-square law here at Birmingham. There were a couple more experimental talks leading into a session on gravitational waves, which I enjoyed particularly. I spoke on a soon-to-be published paper, and Birmingham PhDs Hannah Middleton and Simon Stevenson gave interesting talks on what we could learn about black holes from gravitational waves.

Detecting neutron star–black hole binaries

Slides demonstrating the difficulty of detecting gravitational-wave signals from Alex Nielsen’s talk on searching for neutron star–black hole binaries with gravitational waves. Fortunately we don’t do it by eye (although if you flick between the slides you can notice the difference).

In the afternoon, there were some talks on cosmology (including a nice talk from Maggie Lieu on hierarchical modelling) and on the structure of neutron stars. I was especially pleased to see a talk by Alice Harpole, as she had been one of my students at Cambridge (she was always rather good). The day concluded with some numerical relativity and the latest work generating gravitational-waveform templates (more on that later).

The second day was more theoretical, and somewhat more difficult for me. We had talks on modified gravity and on quantum theories. We had talks on the properties of various spacetimes. Brien Nolan told us that everyone should have a favourite spacetime before going into the details of his: McVittie. That’s not the spacetime around a biscuit, sadly, but could describe a black hole in an expanding Universe, which is almost as cool.

The final talks of the day were from the winners of the Gravitational Physics Group’s Thesis Prize. Anna Heffernan (2014 winner) spoke on the self-force problem. This is important for extreme-mass-ratio systems, such as those we’ll hopefully detect with eLISA. Patricia Schmidt (2105 winner) spoke on including precession in binary black hole waveforms. In general, the spins of black holes won’t be aligned with their orbital angular momentum, causing them to precess. The precession modulates the gravitational waveform, so you need to include this when analysing signals (especially if you want to measure the black holes’ spins). Both talks were excellent and showed how much work had gone into the respective theses.

The meeting closed with the awarding of the best student-talk prize, kindly sponsored by Classical & Quantum Gravity. Runners up were Viraj Sanghai and Umberto Lupo. The winner was Christopher Moore from Cambridge. Chris gave a great talk on how to include uncertainty about your gravitational waveform (which is important if you don’t have all the physics, like precession, accurately included) into your parameter estimation: if your waveform is wrong, you’ll get the wrong answer. We’re currently working on building waveform uncertainty into our parameter-estimation code. Chris showed how you can think about this theoretical uncertainty as another source of noise (in a certain limit).

There was one final talk of the day: Jim Hough gave a public lecture on gravitational-wave detection. I especially enjoyed Jim’s explanation that we need to study gravitational waves to be prepared for the 24th century, and hearing how Joe Weber almost got into a fist fight arguing about his detectors (hopefully we’ll avoid that with LIGO). I hope this talk enthused our audience for the first observations of Advanced LIGO later this year: there were many good questions from the audience and there was considerable interest in our table-top Michelson interferometer afterwards. We had 114 people in the audience (one of the better turn outs for recent outreach activities), which I was delighted with.


We had a fair amount of interest in the meeting. We totalled 81 (registered) participants at the meeting: a few more registered but didn’t make it in the end for various reasons and I suspect a couple of Birmingham people sneaked in without registering.

Looking at the attendance in more detail, we can break down the participants by their career-level. One of the aims of BritGrav is to showcase to research of early-career researchers (PhD students and post-docs), so we ask for this information on the registration form. The proportions are shown in the pie-chart below.

Attendance at BritGrav 15 by career level

Proportion of participants at BritGrav 15 by (self-reported) career level.

PhD students make up the largest chunk; there are a few keen individuals who are yet to start a PhD, and a roughly even split between post-docs and permanent staff. We do need to encourage more senior researchers to come along, even if they are not giving talks, so that they can see the research done by others.

We had a total of 50 talks across the two days (including the two thesis-prize talks); the distribution of talks by career level as shown below.

Talks at BritGrav 15 by career level

Proportion of talks at BritGrav 15 by (self-reported) career level. The majority are by PhD students.

PhDs make up an even larger proportion of talks here, and we see that there are many more talks from post-docs than permanent staff members. This is exactly what we’re aiming for! For comparison, at the first BritGrav Meeting only 26% of talks were by PhD students, and 17% of talks were by post-docs. There’s been a radical change in the distribution of talks, shifting from senior to junior, although the contribution by post-docs ends up about the same.

We can also consider at the proportion of participants from different institutions, which is shown below.

Attendance at BritGrav 15 by institution

Proportion of participants at BritGrav 15 by institution. Birmingham, as host, comes out top.

Here, any UK/Ireland institution which has one or no speakers is lumped together under “Other”, all these institutions had fewer than four participants. It’s good to see that we are attracting some international participants: of those from non-UK/Ireland institutions, two are from the USA and the rest are from Europe (France, Germany, The Netherlands and Slovenia). Birmingham makes up the largest chunk, which probably reflects the convenience. The list of top institutions closely resembles the list of institutions that have hosted a BritGrav. This could show that these are THE places for gravitational research in the UK, or possibly that the best advertising for future BritGravs is having been at an institution in the past (so everyone knows how awesome they are). The distribution of talks by institution roughly traces the number of participants, as shown below.

Talks at BritGrav 15 by institution

Proportion of talks at BritGrav 15 by institution.

Again Birmingham comes top, followed by Queen Mary and Southampton. Both of the thesis-prize talks were from people currently outside the UK/Ireland, even though they studied for their PhDs locally. I think we had a good mix of participants, which is one of factors that contributed to the meeting being successful.

I’m pleased with how well everything went at BritGrav 15, and now I’m looking forward to BritGrav 16, which I will not be organising.