How to Find a Job in Statistics – Advice for Students and Recent Graduates
January 13, 2014 14 Comments
A graduate student in statistics recently asked me for advice on how to find a job in our industry. I’m happy to share my advice about this, and I hope that my advice can help you to find a satisfying job and develop an enjoyable career. My perspectives would be most useful to students and recent graduates because of my similar but unique background; I graduated only 1.5 years ago from my Master’s degree in statistics at the University of Toronto, and I volunteered as a career advisor at Simon Fraser University during my Bachelor’s degree. My advice will reflect my experience in finding a job in Toronto, but you can probably find parallels in your own city.
Most of this post focuses on soft skills that are needed to find any job; I dive specifically into advice for statisticians in the last section. Although the soft skills are general and not specific to statisticians, many employers, veteran statisticians, and professors have told me that students and recent graduates would benefit from the focus on soft skills. Thus, I discuss them first and leave the statistics-specific advice till the end.
Choosing Your City of Study
Most good statistics jobs that I have encountered require at least a Master’s degree. (However, many employers in data science are happy to hire anybody with a strong background in math and computer programming, as well as experience in working with data. This advice post will reflect my experience in working in statistics, but I encourage you to research the job market in data science to broaden your scope.) When students choose where to study for their graduate degree, I think that the industrial activity and job market for statistics in the city of study should be a key factor in your decision. When I was choosing my graduate school for my Master’s degree, I was offered to study with a world-renown statistician who specializes in machine learning, but his university is in a city without a lot of industrial activity for statisticians. I ultimately chose to study in Toronto mainly because of the high quality of the statistics department at the University of Toronto (there are some world-renown professors there), but also because of the high industrial activity for statisticians in that city. (The course-based Master’s program has an intense but short duration of 8 months, which was another attraction for me. Had I wanted to extend my degree by pursuing a thesis or a thesis-like project, the program would have gladly accommodated that, so it was nice to have that option to pursue the more traditional 2-year path with a thesis.) It is harder to find a job in statistics in a smaller or less industrious city like Vancouver, where I currently work. Some of the advice that I offer below would be difficult to implement if you live in a city with little industrial activity or a small job market for statisticians, but you can still do some of it via the Internet through email, Skype and Google Hangouts.
The best way to find a job is via networking. Employers may post jobs online, but they prefer to hire good candidates whom they have personally met in the past in professional settings or via trusted referrals from their colleagues – it’s faster and far less time-consuming than examining hundreds of cover letters and résumés from strangers. I have even encountered some companies that are not actively seeking new employees but will create a new position in order to hire a smart and hard-working person whom they have met through their network.
Networking can be done face-to-face and online, but face-to-face is much better. Your network starts with your professional and academic acquaintances: professors, teaching assistants, classmates, past employers, past co-workers, and colleagues from extra-curricular activities. I did not have a single professional contact in Toronto when I first arrived there, so I had to grow my network from scratch. Attend industrial events to meet professionals in statistics and analytics, and you may be talking to you future employer. This was how I found my first job in statistics after my graduate studies. Be proactive in approaching strangers at these meetings and ask them about their work. If someone’s work especially appeals to your interests, express your feedback and enthusiasm to them, and perhaps even share your thoughts or ideas. (Remember: Networking works best when it’s a win-win interaction. Be prepared to both learn from someone and share what you know.) During my graduate studies in Toronto, the SAS User Groups in Toronto were the best networking events. These meetings are also great ways to improve your skills in using all types of SAS products, from Base SAS to SAS Enterprise Miner. Toronto’s SAS community is especially active, and I attended the meetings held by
The above web pages contain great archives of past presentations about SAS, and they are great learning resources! I also attended some smaller but still very valuable events like
- the Business Analytics Seminar Series held by the Southern Ontario Regional Association (SORA) of the Statistical Society of Canada (SSC)
- the Toronto Applied Biostatistics Association (TABA)
- the Vancouver Machine Learning Meetup group
- the Vancouver Data Science Meetup group
- the Vancouver R Users Meetup group
- the Vancouver SAS Users Group (VanSUG)
Once you meet some professionals or even potential employers, ask them if they would be so kind and available to meet with you for an information interview. These meetings are not about actual job opportunities; they are a chance for you to learn more about that person’s work, company, or industry. Ask them about what they do, what they like/dislike about their job, how they got the job, what traits they look for in ideal candidates in their field, and the current trends in their industries.
Of course, this is an excellent time for you to make a great first impression, because your interviewee may be your next employer or may refer you to your next employer. Be prepared with your questions, be professional in your interaction, respect their time, and express gratitude for their generosity.
Visit your university’s career centre to devise a plan on how you will find a job. (If you are a recent graduate, you may still be able to access those services as a recent alumnus – check your career centre’s rules. I was able to use the University of Toronto’s Career Centre‘s services for 2 years after my graduation.) Meet with a career advisor and be specific about what you want to get out of the appointment. Your first meeting may be to discuss your overall goals and aspirations. The career advisor may point you toward general directions and suggest ways for you to improve upon any weaknesses in your approach or your qualifications. (Even after volunteering as a career advisor for 6 years, I still learned a lot from my appointments with my own career advisor. Those meetings were fruitful because I was quick to identify my weaknesses and prepared for each meeting with specific questions about how to work on my deficiencies.) As you go through the different stages of your job search (searching, applying, interviewing, negotiating, working, moving onto a new job), continue to meet with an advisor to work on all areas of your career development:
- information interviews
- cover letters
- job interviews
- salary negotiations
- researching potential employers
- developing a good online presence
- choosing and preparing your referees
- showing that you are a good fit with your desired potential employer
- phone interviews
Searching for Job Postings
Even the traditional method of looking for job postings has been advanced in recent years – and much of it is due to machine learning! LinkedIn and Indeed are my favourite job search web sites, but there may very well be others that are good. Good job search web sites use recommender systems to automatically find jobs and employers that suit your interests, and they will send regular alerts to you via email about new job postings.
Multiple employers have told me that they value communication and interpersonal skills more than statistical knowledge when they assess job candidates. Most candidates have the basic level of statistical knowledge that they need for the job, but communication and interpersonal skills are much harder to train and teach. If you lack experience with a particular statistical technique, most employers will gladly give you time to learn about it. However, they simply do not have the time or patience to teach you key skills that all statisticians need – for instance,
- explaining a concept clearly to non-statisticians
- writing a report with the statistical analysis plan, the results of the analysis, and their interpretation
- delivering a presentation to a large and diverse audience (i.e. public speaking)
- developing a rapport with people inside and outside of your team
- asking questions to clarify what a client wants to accomplish
- reading a client’s body language to sense repressed confusion or doubt, and taking the initiative to untangle their confusion and instill both clarity and confidence in their understanding
- contributing ideas, asking questions, and even disagreeing with others in meetings, phone calls and teleconferences (very different settings from one-on-one communication)
- writing emails in clear, unambiguous, flowing, professional and grammatically correct language
Thus, one of the most important things that you need to do to get a job in statistics (or any industry) is to develop your communication and interpersonal skills. Regardless of how technical your job may be, you will need to communicate and relate with your boss, co-workers, and clients, and much of that communication involves empathy, active listening, sensitivity, assertiveness, negotiation, compassion, setting boundaries, expressing anger professionally, expressing gratitude professionally, providing constructive feedback, accepting positive and negative criticism, and diplomacy – none of which you will learn in statistics classes. These skills can only be developed through experience, and getting a lot of good work experience or volunteer experience (in statistics or non-statistics roles) is a great way to do so.
Another great way to develop communication skills for statistics is by taking a statistical consulting course, which many graduate programs in statistics offer. These courses will pair you with actual clients who seek your advice and expertise, and they are often non-statisticians who need you to communiate technical concepts in ways that are understandable and practical for them. I took such a course at the University of Toronto during my Master’s degree, and it was a valuable experience for me.
(I offer this following advice as an anglophone working in North America. Please feel free to change the word “English” to whichever language is dominant in your place of work.) Many statistics students do not speak English as their first language. If this is true for you, I highly recommend you to take the time to develop your English skills in social settings outside of statistics and away from the classroom. Read properly written English closely and learn to mimic correct grammar until you grasp the rules yourself. Train your accent to become closer to that of your native English-speaking peers. I have met many smart and hard-working statistics students whose talents are not fully realized in the working world because they are hesitant to socialize with people who don’t speak their native tongue; while this is easier for making friends, it also greatly limits your ability to develop your English skills and familiarize with the local culture. Try to socialize with anglophone people; a friendly person with mature wisdom and a kind heart will shine through any language barrier with like-minded and like-hearted people.
Advice Specifically for Statisticians
- Learn both SAS and R. Academic statistics tends to use R, but many statistics jobs require good knowledge of SAS. There are many good resources to learn both languages online. My blog offers many resources on R programming, and I’m beginning to blog about SAS programming, too.
- Learn machine learning. Regardless of which field you are working in, machine learning has already become prolific in all areas of industry, and will become even more prolific in the near future. If you haven’t learned any machine learning in school, have no fear – there are some great resources online for free! Many massive open online courses (MOOCs), like Coursera, teach machine learning, and most include programming-intensive projects and assignments that you can later display to your potential employers as evidence of your skills. There is a great Youtube channel by Mathematical Monk on machine learning that I also like.
- Most biostatistics jobs that I have seen require knowledge of survival analysis as a basic requirement. They also require knowledge of how to use survival analysis in SAS. I have really enjoyed reading Paul Allison’s textbook, Survival Analysis Using SAS: A Practical Guide, to learn both.
- Despite the emphasis on linear regression for modelling continuous response variables in statistics education, I have found that logistic regression is used a lot more often in industry. (I have worked in industrial statistics and medical statistics so far). Thus, develop a thorough understanding of all aspects of binary variable analysis: 2-by-2 contingency tables, specificity, sensitivity, positive predictive value, negative predictive value, concordance statistics, ROC curves, chi-squared and Fisher’s exact tests of independence, and – most importantly – logistic regression.
- In biostatistics, knowledge of statistical genetics is becoming increasingly more valuable. If you want to work in biostatistics, being good at it will certainly give you an edge.
- The American Statistical Association has a great web page with data on salaries that you can use for your salary negotiation.
- Take the initiative to show samples of your work, even if most employers don’t explicitly ask for them. It’s a great way for you to concretely demonstrate your technical and theoretical knowledge. For example, I once brought a MATLAB script implementing a recommender system to a company to demonstrate my passion for machine learning, collaborative filtering, and computer programming. I also brought two 1-page descriptions of what I did: one in words, and one using a flow chart.
- An even better way to show your work is to write a blog. I will write another advice blog post about that; in the meanwhile, check out my blog or the blogs in my blog roll for some good examples.
- When I was a student, many of my professors told me that employers will generally tolerate job applicants not having certain skills like data manipulation, SAS programming, or programming in some other language (like Python, SQL, C++ or Java) – skills that are they considered to be relatively easy to acquire with 1-2 months of “learning on the job”. My experience in the job market has taught me that their advice has been mostly wrong. Most employers who demand those skills will not consider candidates who lack them, even if they are relatively strong in their statistical knowledge. They have work for their new employees to do right away, and they usually do not have the patience for them to learn those essential skills (especially computer programming in a particular language) on the job. (Some employers may give you some time to develop your skills in data manipulation for a little while, just because you need some time to learn about the data and become familiar with their intricacies and quirks.) Plus, there are many people who do have those skills, so, even if an employer likes you, other candidates may simply out-compete you because of their larger skill sets. Thus, at least a year before you expect to start working, identify the essential skills in your desired jobs and acquire them while you are still in school. Use your spare time – especially your vacations – to acquire those skills while you are still a student. By the time you start applying for jobs, you will be prepared to serve employers with skills that they need, even if your school doesn’t teach them.
- In many statistics jobs, including my current job at the British Columbia Cancer Agency and my previous job at the British Columbia Centre for Excellence in HIV/AIDS, much of my work involved cleaning raw data, manipulating data into the appropriate formats, and merging different data sets to obtain the final data set that I need. Only after hours or even days of doing so can I even begin to use my statistical knowledge to analyze the data. Unfortunately, my academic education in statistics did not teach me skills in data manipulation and processing – the data sets in my homework were already cleaned and formatted for analysis. Some companies – like Predictum, where I first worked after graduating from my Master’s degree in statistics – compartmentalize these tasks so that the statisticians can focus purely on methodological research and analysis. Predictum’s database managers and clients usually provided the data in the format that I needed, so I did not have to do any data manipulation. I liked that a lot, but I have learned that this compartmentalization is the exception rather than the rule. Thus, I highly recommend you to learn how to manipulate data in both R and SAS. I just provided links to 2 resources that I found online – please share your favourite resources in the comments below!