Overcoming your peers fears
When I started my PhD, I felt incredibly self-conscious that I was starting a doctoral program in social robotics without ever having written a line of code. I thought I would have to fight to convince my peers to accept me as an equal, primed with years of hearing that psychology is a ‘soft’ science, less valid than engineering, computer science, maths or physics. Even now, receiving emails targeted at ‘women in STEM’ fills me with a tiny pang of guilt. Imposter syndrome is no stranger to many PhD candidates, and working among peers with completely different skillsets and backgrounds to you can amplify those feelings just a little more.
The good news is, all of these fears turned out to be unfounded - my fellow PhD students in engineering and computer science have never once suggested that psychology isn’t a real science, that their degrees are harder, or that I don’t deserve to be here. Instead many have approached me with questions about research design and statistics, with a genuine interest in bettering their research through psychology. Such interactions not only increased my confidence in my own knowledge, but highlighted to me the importance of being being open-minded when approaching peers from different disciplines, and to beware of not only what you think others’ biases might be, but what your own are as well.
Interdisciplinary Love Languages
Like any partnership, interdisciplinary projects are not without their bumps in the road. Collaborations can often lead to disagreements over sample sizes, research methods, and do we really need to submit an ethics proposal? However, just as my collaborators sometimes struggle to understand my insistence that we need more than 20 participants, I don’t always understand the tech behind human-robot-interaction scenarios. Words and phrases like machine learning, Wizard of Oz, latex, APIs, IDEs, SDKs, conda, jupyter, numpy and so many more can often be scary and overwhelming. At some point though, I realised that as unfamiliar as these things seem to me, so did things like power analyses, confounds, ANOVA, pre-registration, and HARK-ing seem to them. I have endless appreciation for my many colleagues who are patient and understanding in explaining things to me as I continue trying to navigate the world of robots (what is github? What is overleaf? What is VIM? How do I exit VIM?) In turn, I explain how to design experiments; what are independent and dependent variables, how to calculate sample size and power, how to write an ethics proposal. It is a process of not just reciprocal teaching but learning how to communicate. Things which seem so completely obvious have to be completely broken down to find common ground and any assumptions of ‘basic’ knowledge or ‘everyone should know that’ thrown out the window - on both sides. This process, of learning not just what to communicate, but also how to communicate is one of the most valuable skills I have learned throughout my PhD so far.
Reverse Discipline Shock
The expats among you might have heard the term ‘reverse culture shock’, where when you return home after travelling or living abroad there is a period of readjustment. Interdisciplinary projects can be similar - returning to your ‘home’ discipline after spending time away in the labs of robots can be jarring. You have just learned to communicate with your non-psychology peers, now, however, you might feel you somehow relate a little less to non-robotics psychologists.
My first experience with this came about one year into my PhD, when I attended a conference targeted at open science in psychology. After having spent the majority of my time so far around computer scientists and engineers, I was excited to be among psychologists again. Here were people at the forefront of the open science movement, proactive about pre-registration, preprints, open data, and open access. Alongside this were also huge discussions about statistical techniques, from data scraping to big data to multi-level modelling. All these things are great and necessary for the progression of psychology as a field. Yet, as someone working in a field that is not pure psychology, I suddenly felt out of place. How can you promote pre-registration when your collaborators don’t know what an independent and dependent variable are? How do you run sufficiently powered studies in a field which commonly accepts and publishes with N < 30? I felt not only that there was no room for discussion of gray areas, but that I was out fighting my own battles for psychology on a completely different battlefield. This is a dichotomy I am still trying to reconcile - keeping up to date with best research practices in psychology whilst designing HRI experiments can often be a balancing act.
So, to avoid falling over too often, I try and keep the philosphy of work with what you have, and be transparent about what you did, and why. No study is ever likely to be perfect, even moreso when you add in robots, but that does not necessarily mean the contribution is not valuable, or that the findings should not be shared. It is possible to both highlight the contribution of a study whilst still acknowledging it’s limitations. If you try and fit into the box of entirely psychology or entirely HRI, you will end up pleasing neither. Instead, try to embrace the creativity that comes from not being bound to the confines of one field. It might feel like a weakness at times (psych venues might reject your work for being too niche, HRI says its too psych-oriented), but to me, being able to develop your own unique perspective informed by multiple different disciplines is one of the true strengths that comes out of working in this field.
Resolving the Psydentity Crisis
Nearing the end of my PhD, I now feel more comfortable about the role myself and other psychologists have to play in social robotics. An important part of this journey was recognising my own strengths and weaknesses, as well as those of my collaborators, and what we are each contributing to a project. Someone at a conference recently told me ‘creating an interdisciplinary team for human robot interaction is like creating a startup’ based on the number of different skills needed. And it is true - whilst there may be some unicorns out there who can write code, design interactions, run flawless statistical analyses, and know every relevant paper ever written on a subject, the reality is, most of us are not like that. Although we should always strive towards learning, trying to become a master in each of those domains (especially during a PhD) is a recipe for a lot of frustration. At some point I realised that the goal should not be for me to try and become an engineer (or computer scientist, or UX designer), nor for my colleagues to become psychologists. Instead, we aim to build upon and complement each other’s skill sets. The end result is studies and collaborations which go beyond the limits of what any one field or discipline can achieve, which to me, is the essence of being an interdisciplinary scientist.