Embracing Imposter Syndrome
Recently I was referred to by a colleague as the resident data and engineering “expert.” I was initially taken aback; I’ve never considered myself much of an expert in anything, let alone this portion of my career I’ve largely taught myself. With this referral came a mix of feelings including fear and ire. Fear came from self-doubt of living up to expectations, ire from blaming my colleague for positioning me precariously to potentially fail. Why did I feel this way? And why have similar feelings come up so frequently throughout my adult life? After all, I’ve had a pretty good track record of not falling completely flat on my face when tasked to solve a problem.
The fear of being “found out” is common for many people, especially younger and newly experienced workers. It’s even more prevalent among women, people of color, and underrepresented populations. There is no job posting that doesn’t expect a new hire to “hit the ground running,” take an enterprise to the “next level,” and bring fresh “innovative” ideas to a team. Do you recall the last time you heard a job advertised at a leisurely pace to set you up for success? The onus lies on the person holding the position to demonstrate that they are capable. Not that this is a totally bad thing, it certainly wouldn’t benefit to incentivize laziness or remove accountability. However, in the hyper-competitive job market of the modern era, as is especially true in the world of data science, it’s incredibly difficult to dismiss an innate need to constantly and perpetually prove one’s worth for any amount of security and satisfaction. The last thing anyone wants is to be branded as “expendable.” So many of us label ourselves “imposters.” A mix of guilt for taking merit we don’t feel like we deserve, fear of confirmation for being less than we advertised, and anxiety over the maintenance of a shaky state of being.
Imposter Syndrome in Relation to Data Science
The special brand of this syndrome that I have experienced, and am experiencing, I feel is especially poignant in the data science and analytics field and I don’t think I’m alone nor do I think it’s coincidental. “Big data” in all its popularity only came around as recently as the 1990’s, and “data science,” as a mix of disciplines from computer science and statistics, wasn’t in the public lexicon until the early 2000’s. When I graduated from university in 2015, I couldn’t name a single school offering a degree in data science. The field itself is new, but the demand for it has never been higher. Organizations of every type and variety are all clamoring over the need to analyze their data for quantifiable gain. Hospitals want to predict severe illness and improve patient outcomes, sporting teams want to seize the greatest probability of how best to overcome the opposing team, big businesses want to monitor customer demand and satisfaction. The list is infinite. But hardly anywhere has a standardized, dedicated track to create the true, perfect “data scientist” like traditional hard disciplines of mathematics, engineering, and biology. Is it any wonder that the individuals in this field would feel a little shaky when the field itself is still forming?
Analysts, programmers, and engineers are often tasked with making a product that solves a problem. Whereas in traditional structural engineering you might picture a bridge or a vehicle, in data science that product will often inform decision making or optimize a particular workflow. The responsibility lies on the architect of that solution to produce results that are repeatable, reproducible, and reliable (three core tenants of engineering and statistics). Just like the bridge or the car, the user of your product carries a level of trust and the impact of it can be immense.
When I majored in mechanical engineering for undergrad, one of my internships was with a small company that contracted to Boeing. During my 6 month time at the internship I worked on, in hindsight, extraordinarily insignificant parts of the 747’s door frame that absolutely never even came close to making it on an actual aircraft (not that I’m upset, I wouldn’t want to board an plane that a 19 year old had slapped together either). However, I remember feeling pressure even then to force myself into a position of ownership over something that was much greater than I felt capable of claiming. A plane? Even if I was just making a box to store the fizzy drinks it seemed crazy to work on something that would shepherd countless passengers through the sky.
Fast forward through a few more internships and positions and I still experience the same tumult in stacking myself up in the field of my colleagues. When I first stepped figurative foot into the field, I was constantly making one step forward with what felt like a giant leap back. I picked up R first, but how confident could I be in it with no foundation in traditional computer science? Do I take up C? Is R correct, should I learn Python, too? What about that whole machine learning and AI thing that’s making such a boom… I haven’t taken a stats course since high school!
For about a month or two I would spend my evenings after work trying to teach myself more skills that would generate value at my job. Thankfully to some degree it worked and I now have numerous workflows successfully running to support my various projects. Yet there’s no denying how bumpy, disorganized, and wholeheartedly unique that starting point was for me and I’m sure is for many. There have been many discouraging moments that may or may not resonate with people from or in similar situations:
- I have no PhD
- My majors were not in a traditional statistical field nor were they dedicated computer science
- Many of my colleagues all do have that traditional foundational background
- Stack Overflow has at times not been the most kind to my earnest questions
Overcoming, Embracing, and Persevering
For what it’s worth, if you’re reading this and feeling disheartened or worried or scared understand that you’re not alone and that in my own unique way I get it. Prior to working at CHOP, I had much less job security and feared being unemployed at a moment’s notice. Regardless of my level of passion and commitment to a role, with my own student loans, expenses, and bills to worry about there has never been room to fail at a job. Along the way to feeling a bit more at ease, while I’ve never truly gotten rid of the feeling of “faking it till making it,” I’ve made a few observations that have helped.
Everyone Starts at a Different Position
So you didn’t get a PhD, or only just discovered your passion for machine learning, or all those years that you thought you hated coding were flipped upside down when something “clicked” and you embraced a particular language. Those are all ok. Everyone has a different starting position with a different set of circumstances that bring them to where they are, and not just in data science. While the often-self-imposed isolation felt from being out of place in a role can be paralyzing, it is not insurmountable. Your ceiling for value generation and potential is not capped by past milestones, rather it is dependent on your motivation and drive to equip yourself for success. There is no wrong time to make yourself useful and valuable to those around you.
That being said, I would be remiss to not acknowledge that getting to a desired position in a career, or in life is disproportionately, weighted in favor of specific people, namely those who are white and, especially in data science and engineering, those who are male. I immediately think of the Life of Privilege Race video which many have seen (if you haven’t I strongly recommend clicking and checking out), which distills this metaphor of starting positions into very literal realistic ones. I cannot, as a white male in this field, say that I can give expert advice on how to overcome these obstacles that I do not face. But I encourage those of you who do face these obstacles to not lose hope to take away only what you’re comfortable with from my limited world view.
You Versus Them
I have friends who do theater, art, dance and are much more right-brained than I am. My audience exposure has been more limited to lectures, presentations, and webinars. Whenever I’ve presented my work to a group or discussed projects with peers I deemed my superiors, I’ve encountered the same fear and doubt. This person is far more experienced than I am, who am I to tell them about the subject matter? That person is an expert in a different field, what if they deduce that I’m not as proficient as I claim in mine?
I’ve made a few observations that I’ll boil down to Rich’s 3 Laws of Very Bigly Huge Differences:
1) There is a very bigly huge difference between what you think you know and what you actually know
2) There is a very bigly huge difference between what you think others know and what they actually know
3) There is a very bigly huge difference between how you think others perceive you and how they actually perceive you
In all of these observations there is a pseudo-quantifiable gulf between expectation and reality. On one side is the theoretical trivial amount of value you bring to the table compared to the goliath machination of expertise and wisdom you make out of the person before you. But the other side hides behind the curtain of doubt and self-restriction. More often than not when I thought someone else knew more than me, would belittle me, or might distrust my input that I actually received the opposite. Often times people were kind or actually found great value from my insight even on matters where they already knew the subject material. Other times I mistook that I actually did know enough to teach someone, and in fact knew more on the material than I thought I did. Granted this won’t be the case every single time. Surely there will be times where you are not the domain expert equipped to deftly deflect every question like a master ninja. But the times where you will be made a fool of for it are much fewer and further between than television or your imagination portray.
Resolution and Embrace
People close to me know that I am a serial overthinker and a bit of a worry wart (they’re probably shaking their heads right now). Admittedly this isn’t the correct way to go through life and usually adds much unneeded stress to my day, but leveraging that stress in a productive way to fuel motivation can help. If I’ve found myself incapable of shaking the feeling of imposter syndrome, what better way to resolve that turmoil than to push myself to becoming the expert I want to see?
When I was in engineering school, I found there was a distinct difference in the study habits between people in my class and folks from the more qualitative fields such as biology and medicine (I won’t begin to claim any expertise in other fantastic fields like literature, language, art, history, etc.). Engineering was a constant application of mathematics and physics; you could study for as long as you wanted but you were never going to know what the problem on the exam was going to look like. You couldn’t memorize a textbook and hope that the knowledge to answer exam questions lay somewhere in your memory. This applied studying was like working out a muscle: it required training and practice and at some point, the anxiety over the next day’s exam reached a key saturation point where no further preparation would yield any benefit.
That key saturation point is what helps to throw away the feeling of imposter syndrome. Put in the work, afford yourself the diligence to rise to the occasion, and the confidence to handle the expectations of your peers. Then when you take the stage you’ll be surprised and elated to see just how well you perform.
For those of you who experience imposter syndrome on a rolling basis, I hope this has been something you could relate to in some capacity and if nothing else that you can walk away with a little bit more confidence in yourself to excel in your current situation. If you are in data science, I hope you are able to alleviate some of that self-imposed burden due to the nature of the field. Whatever your circumstances, you’re going to do great.