A couple of months(?) weeks(?) ago, a colleague at work finished writing up a 1-pager overview of the Quantitative UX Researcher role at Google, available here https://g.co/GoogleQuantUXR. I’ve sorta been too distracted by other topics to get around to writing about qUXR until today.
As a reminder, while I indirectly mention colleagues and various peers within the UX research community, I’m only writing about my own experiences and thoughts here. My previous posts about what qUXRs are, and how becoming one is tricky are also available.
When I look around at my fellow Quantitative UX Researcher colleagues, and also when I’m asked to interview potential candidates for new quants, one thing that is noteworthy is that a majority of them have PhDs. I “merely” have a Master’s degree, and I’ve heard that a small number have Bachelor’s degrees.
This isn’t going to be a post about imposter syndrome or similar themes. There’s never been a point where anyone’s degree mattered, or even came up in the conversation except in the context of quants surveying quants and understanding what our small discipline looks like. Many of us are 5, 10, 15+ years into our careers already and that’s well past the point of where a degree means anything. Plus, I’m cocksure enough about myself to not worry about this.
But the thing that I wanted to think about, and thus write about, is what sorts of paths can someone without a PhD, or any graduate training whatsoever, take to becoming a quant.
Why PhD’s tend to have an easier time becoming an qUXR
The most important part of being a Quantitative UX Researcher isn’t the quantitative part, it’s the Researcher part. The whole point is to understand user experience for some business purpose. In an ideal world, there’s no specialization between quantitative and qualitative researchers — we’d all just be researchers. Methods should always follow the research question, regardless of whether this is in industry or academia. Something might have gone horribly wrong if you’re spinning up a hadoop cluster to (somehow) analyze 5 user interviews.
The only reason the UX Researcher title has started to separate into two groups is because it’s very difficult to find anyone who can both quant and qual methods at a high level. We all wind up leaning towards one or the other, even if we have experience using both.
Data scientists, which shares a lot of skills as qUXRs, are already facing some pressure to break into sub-fields like ML Engineer, Data Engineer, or Analytics Engineer because the volume of knowledge is growing too large for a single human to master. It’d be madness to say we want full fledged data scientists who can also do qualitative research.
Since the core of the function is “researcher”, people with graduate school experience, whether PhDs and certain Master degrees (typically the ones that involve writing and defending a thesis) tend to have an advantage because those degree programs are effectively apprenticeships for learning how to do research. Both quant and qual UX researchers populations have a large bias to PhD folk for this reason.
Being a researcher is more than simply thinking up a research question and executing the technical details about it. Training to join academia involves learning about the contents and edges of a field, the accepted standards of proof and writing, the existing community and conversations of the field’s members, knowing how any project fits into that community. It’s a complicated craft that’s most commonly taught by a mix of reading and analyzing research papers, doing hands-on research, and also learning by example through watching and being advised by existing, experienced, researchers.
While that list of stuff sounds like abstract nonsense when considering the day-to-day work, it’s foundational stuff. It affects how you pick research questions, how you select methods, and just how you approach things in general. The many PhDs in the quant community at large in the world bring this foundation into the discipline and collectively it sets a baseline assumption of how work “looks and feels”. This is true even though we’re not in academia any more and are serving the less rigorous, faster paced needs of industry. Even if we all come from different fields with like statistics, social science, neuroscience, that can differ quite a bit in terms of community and methods, there’s more shared than not.
But the point I want to strongly emphasize here that I, and many of my colleagues, don’t believe that the job specifically demands a graduate school degree, or any degree. It’s actually a detriment if we exclude extremely smart and talented researchers simply because they didn’t opt to go through the “traditional” academic route. If we started getting lots more applicants from the data science community at large, which shows a more normal degree distribution (though still highly educated), we’d be overjoyed.
So what is someone who doesn’t have access to a graduate degree to do? Academics don’t have a monopoly on teaching/learning how to do research. Apprenticeships are merely controlled ways to get hands-on experience and guidance. It can’t be the only path, there must be alternate ways.
While I do come from a semi academic background, I can see a couple of viable pathways.
Idea 1: Finding work within UX research directly and climb up
Large research teams are typically organizations within themselves. While there’d obviously be a lot of researchers, there are usually various levels of analysts and assistants involved to keep things moving along. Titles can vary wildly, but you’d want to keep an eye out for things like research assistant, research analyst, UX analyst. The key is to look for things within the a UX research organization.
Some of these positions are designed to be entry level, others might only require a relatively small amount of experience. Either way, these position are great because you’ll be asked to help more senior researchers execute research projects. Expect to be hands on cleaning survey responses, coding text or behavior, analyzing video, scheduling sessions, recruiting users, etc.. When you make mistakes, they should be able catch most of those errors and correct you.
On top of the direct execution of many aspects of research, such positions would also be either involved in, or bear witness to, the whole process involved in planning and negotiating research. Those processes are generally very messy and hard to explain clearly in words since it all involves juggling various factors.
Finally, such positions will let people experience research processes. How do teams handle questions coming in, do they use tickets, how do they push back or redirect questions, how to they figure out what’s worth doing. Good processes are always important, and they’re often a form of long term institutional memory that you rarely get access to seeing on the outside.
Idea 2: Get into analytics and product work and climb up
Even without explicitly doing “research”, spending years working and thinking about how to measure, analyze, and improve products will hone most of the same skills. Ideally, this is done under the supervision of someone that does have significant experience in the particular domain so that there’s guidance on what methods to use and pitfalls to watch out for.
These positions might be called various forms of “analyst”, data analyst and business analyst are the most common, though you’ll also find examples under things like “Business Intelligence (BI)”, or under the umbrella of “product development” in certain industries.
The great thing about this way is that starting out from an analyst-type position is that there are lots of potential routes to choose from as you gain experience. Many data scientists come out of analyst backgrounds (myself included), while others may branch off into management, research, data engineering, ML, or research. All of these choices are completely just based on personal interest and aptitude. You’re not committed to the whole “UX Research” thing yet.
You can see some examples of massive title variation for many data scientists in the thread below:
Whatever you choose, you’re going to wind up developing a lot of valuable domain knowledge as you work. This is actually an advantage when compared to people who come out of academia because someone fresh out of a sociology program is very unlikely to have strong knowledge about whatever industry they’re joining. Sure, they can apply their research skills to approach any problem and ramp up, but adjusting can often take half a year or more of work on their part.
Idea 3: Get into an adjacent position and transfer laterally
Product work and UX work all involve tons of functions. You’ve got engineers, designers, writers, lawyers, business folk, all working together. Many of those functions wind up consuming whatever output the researchers generate. They’re the research team’s stakeholders. Find a way to become one of those people.
Over time, research skills can rub off, if you let it. I know some people who started their careers in things like design or engineering, but after working with researchers for years, they developed an interest in the field, built up their skills, and made a lateral transfer to become a researcher of some sort. It obviously took a lot of extra time and effort on their part to do things like learn how to run a research study themselves with advice from the research team. But it does work.
Idea 4: Do any of the above, and make a point of taking a research methods course or two
The amusing thing about graduate school is that, unless you’re specifically in a program that’s about doing research in statistics or research methods, everyone pretty much takes just one or two general purpose research methods courses as a base foundation. Afterwards, they might take a specialized course for whatever state-of-the-art method they’re trying to use in their research. Luckily, thanks to publicly available course resources, you can probably take some of the equivalent.
The biggest decision to be made is picking which field’s research method's course do you want to take, because every field varies on what they find acceptable. While a survey is a survey the world around, occasionally, terminology and names of more obscure methods can even differ between fields. “Research methods for engineers” courses are usually quite different than similarly named courses for sociologists, economists, neuroscientists, etc.
My personal bias and preference is choosing methods classes from one of the social sciences - economics/econometrics, computational sociology/linguistics, psychology, political science, communication, etc.. The methods themselves can range from being very mathy and technical (econometrics is often like this) to being more about classical survey methodology.
The social sciences also have issues and long internal debates over whether we’re really measuring abstract theoretical concepts like “happiness”, and understanding ways to operationalize and validate measures. They also have an identify issue of needing to justify that they are a “Science” and so they include brief lectures on philosophy of science which I find useful to know. The constant lingering doubt about everything in the back of your head is very useful.
As a bonus, because most social science effect sizes tend to be tiny (as are the sample sizes), you get exposed to methods that are useful in murky situations. You’ll be dealing with all of these issues in your daily work, so it’s all very relevant.
I’ve tried searching online for good examples of such courses, but it’s a confusing mess. This particular course on EdX from the National University of Singapore resembles what you’d want to look for. It’s got dedicated sections on formulating research questions, doing lit review, and ethics on top of experimental design and a wide mix of qualitative and quantitative methods.
The goal is to get a passing familiarity with a wide array of methods first so that when a research question comes up, you’ll have a toolbox to select from. Once you know what specific family of methods you want, you can put the energy into learning to use the particular method in detail.
I also don’t have any good books to recommend (if someone has a favorite methods book, let me know). I’ve scanned a bunch of syllabi for methods courses and I don’t think I’ve seen the same book twice. There’s no agreement that I can see. You’re probably better served looking for popular/highly reviewed introductory methods textbooks on Amazon.
Relax, there’s no one right way
Probably the only thing I can take away from scanning a bunch of syllabi and methods book descriptions is this — there’s no agreed upon way to become a researcher.
Even if you go through the academic route, we all learned a bunch of random methods in our methods courses, of which we use only a handful repeatedly. All the books differ, all the fields care about different stuff. The fact that such a motley crew of people coming out of all these programs can come together and define an industry researcher position tells the story that the specifics don’t matter. If you’re smart enough to figure out how to do original research in one setting, you’ll be able to figure out how to do it in other settings.
About this newsletter
I’m Randy Au, currently a Quantitative UX researcher, former data analyst, and general-purpose data and tech nerd. The Counting Stuff newsletter is a weekly data/tech blog about the less-than-sexy aspects about data science, UX research and tech. With occasional excursions into other fun topics.
All photos/drawings used are taken/created by Randy unless otherwise noted.
Supporting this newsletter:
This newsletter is free, share it with your friends without guilt! But if you like the content and want to send some love, here’s some options:
Tweet me - Comments and questions are always welcome, they often inspire new posts
A small one-time donation at Ko-fi - Thanks to the folks who occasionally send a donation! I see the comments and read each one. I haven’t figured out a polite way of responding yet because distributed systems are hard. But it’s very appreciated!!!
I designed a silly mug last week because of a silly twitter tread about taking production down