Breaking into Industry as a Social Scientist
Stop envying natural sciences and market yourself better
I graduated from a social sciences institution. I wasn’t very high on academia—stemming from a combination of having worked in private sector before and not being an idealist—and towards the end of my PhD, I was on the lookout for career opportunities in tech.
I was technically adept; I had been teaching game theory and statistics for several years, I utilised a variety of machine learning models in my PhD, and I knew my way around developing libraries in R, version control, UNIX shell etc.
But it was difficult to get institutional support. Career-wise, as a computational social scientist outside of academia, the avenues available to me via my institution were finance and management consulting. People kept telling me why I don’t just interview at hedge funds and quantitative trading firms 🤷🏻
Turns out, going into tech is a natural sciences thing.
And it made sense for a brief second.
Then, not really.
Are you telling me, for a majority of industry jobs, natural sciences > social sciences?
That doesn’t hold up at all.
Now, I also don’t think one needs a regular/graduate degree or has to attend a pedigree institution to land a data science/machine learning/‘AI’ job. But i) most hiring managers erroneously think so, and/or ii) the market is over-saturated so they can get away with it. But that’s another post for another audience.
The Fantasy of Natural Sciences
If you are quantitative/computational social scientist, there is this envy of natural scientists. You want to be respected like a mathematician, a physicist. Given that there isn’t much to be envious of their social skills and grace (observe mathematicians interact in their environment), I guess it is a combination of how we historically celebrate the accomplishments of natural scientists and how they are/or perceived to be more numbers-savvy (hence, more science-y).
I know because I had a bad case of it!
In my case, the yearning was more about the ability to discover laws. The possibility to uncover regularities that can be expressed with elegant formulas and Greek letters and typeset them in LaTeX. The holy grail of the Western scientific thinking—the scientific method and experiments—we chase that, because we think it will give us legitimacy.
In academia, you often hear about how social scientists bring shame to the cult that is quant—the replication crisis, several high-profile cases of fraud, high rates of p-hacking including both the intentional (low-profile, mundane fraud) and the unintentional (not understanding the methods used/inability to follow procedure) kind.
These are all true. Social scientists are crappy statisticians. They want to do social science, and when their n is high, they borrow a book and read a chapter or two.
[To be fair, the replication crisis exists in other fields (Medicine👋) and p-hacking/mundane scientific fraud is a by-product of the publish-or-perish paradigm, so they are not specific to the social sciences.]
So, what’s the rationale for hiring a PhD in theoretical physics if you are, say, a market research company?
Beats me too.
Once in an interview where I was part of the hiring team, a physics PhD explained to the panel what an Euclidian vector is for a good 30 mins:
Candidate: You know, they have magnitude and a direction…
Panel: Yes, we know what vectors are. Given our industry—market research—what skills would you say you would bring to the team?
Candidate: Ehm, so vectors…
One safe guess would be that we think physicists = smart, PhD = smart, so it’s a quite strong signal of being smart? Or, riding on the natural sciences number-savvy hype (although theoretical != maths), they must be good quantitative researchers who can code?
I think it’s simply hiring managers i) not knowing the skillset they need for their industry (because they were hired haphazardly themselves by a random CxO), and ii) having a natural sciences background themselves. Then, the very same bias perpetuates the system—they just hire more of their clones because someone happened to hire them first.
Why Social Science is a Better Fit for Most Industry Jobs
This probably sounds like a bad case of copium, but I encourage you to keep reading.
We are in the business of data. Most companies generate, store, and track human data (e.g. purchases, behaviours). Human data is messy, noisy, unreliable—it’s a nightmare. It’s not neat as orbits and velocities that one can calculate precisely, over and over again, at different times and places.
Humans, on the other hand, behave in unpredictable ways and belly-laugh at your research designs:
Oh yeah we had 10 control variables in place but the subject had a fight with their partner in the morning and the barista made an error with their coffee on the way and BTW all the data are self-reported
Discovering this kind of underlying data-generating process is the worst day of a natural scientist’s life. For a social scientist, it’s just Tuesday—and by Friday, they will also find out that the random sampling was not—oops!—random.
In industry—well, generally in life—you want people who are comfortable with uncertainty, especially when it comes to decision-making. Now, social scientists often are not directly trained in this, but more so as a function of doing social science and having to deal with crappy, self-reported human data (survey data 🤮). It’s merely part of the job, no biggie.
You should highlight this divine skill of yours! This should be your motto. You are used to crappy data. Natural scientists, they merely adopted it. You were born in it, moulded by it (Chill out, Bane). And you know how to work with it, because you never had good data in your life. You wouldn’t know it if it hit you in the face.
Now, if you are working in a robotics company, go hire robotics people—it helps. But most industries—and I mean the vast majority—are old-school businesses that may or may not need to be DS/ML/AI-enabled. The leaders probably say they want to, but that’s investor pressure and FOMO. It has been ten plus years since DJ Patil coined the term data scientist; what percentage of companies got a positive ROI from their data science investment since then? Hint: not much.
Most companies have terrible data. Spreadsheets-as-databases. No data contracts. Excel exports. Extrapolated numbers (by dragging them down!) beyond their logical ranges. Welcome to the industry!
Social Scientists Shoot Themselves in the Foot
Scientists, in general, suck at communication—just look at the narratives surrounding climate change—with all that evidence, you’d think we would be at a better place when it comes to public opinion.
Some scientists are better at this than others sure, but even those might struggle with translating their work for non-academic audiences.
Social scientists, thinking that’s not enough of a challenge, elect to further sabotage themselves when applying to industry jobs by writing the worst resumes possible:
Name Surname, PhD
PhD thesis that has no connection whatsoever to the job
Work experience: GTA, TA, invigilator (?!)
Highlight: Taught classes in unrelated topic to the job
Achievements: Papers, conferences, awards
No mention of practical skills
Effectively apologising for wasting the time of the hiring manager
These kind of resumes, mostly following this structure out of ignorance, paints you as someone who
Identifies as an academic
Which is a net-negative in industry, because the perception is that academics are slow-paced, ivory tower perfectionists whose only replies to every business question are it depends and more research needed.
Doesn’t understand the needs of the industry they are applying to
Even if you had no life outside academia, you should still make an effort to understand how that specific industry works. Not applying to two million jobs helps with this one.
Is unable to translate their work to showcase generalisable skills
A broader case of the above, you clearly demonstrate that you are not quite capable of translating your past experiences in a way that the hiring manager can assess. When you fail to explain something, people don’t tend to think they are stupid; they think you are—that’s life.
Has this weird tendency to use the exact titles of their graduate work
You were a GTA but your work consisted of data cleaning, model fitting, and parameter tuning? Call that gig data scientist on your resume. When you say GTA, people don’t know what it is and/or don’t care. When you say you were a data scientist, they may still not care but at least they can imagine the type of work you did.
It doesn’t have to be this way! Perceive your job application from the perspective of a hiring manager. They have a need and a budget (headcount), and they want to hire someone ASAP. And I mean ASAP—no one likes interviewing candidates as a pastime. Other than perhaps super-competitive big tech gigs, most companies hire the first good fit and somewhat competent person they can find.
Instead, reframe your academic journey in a way that makes you marketable for industry.
You have a PhD—great! Forget about the degree. Identify what skills were required to complete a dissertation. Perhaps you can meet hard deadlines. Work well autonomously. Work well in a team (if you had a lab). Provide and receive constructive feedback (don’t be a reviewer #2). You are simply resourceful.
You taught classes—great! You are good at public speaking/performance. You can distil the essence of complex phenomena. You excel at answering questions/being put on the spot. You are comfortable with saying “I don’t know”.
You are published—great! You are great at understanding and meeting specs/follow instructions rigorously. You can take belligerent feedback and survive. You are highly resilient.
Make it about what you bring to the table, in a language that the other person can appreciate. Don’t just say you have a PhD, you published research etc.—no one cares. Demonstrate that you are a problem solver, you understand working with constraints, you don’t need babysitting, you meet deadlines. The chances are, they will be paying you for solving mundane, day-to-day problems. You had it way worse. Don’t be apologetic regarding your academic background; turn it around and make it your main marketing schtick.