Blog Viewer

Preparing New Grads for Employment

  

As part of our section's commitment to the International Year of Statistics, I visited several classrooms across Michigan -- and, one job fair at a local four-year institution.

This university has a well-regarded professional science master's program in biostatistics, with an emphasis on health. Because my day job is with a major health care organization, we attended this university's job fair a few weeks ago. We (me, my boss, a co-worker and one of our H.R. folks) spent several hours talking to soon-to-be grads from the undergrad STEM majors as well as the M.S. biostats folks.

Some takeaways:

  • A lot of these young people know NOTHING about how to comport themselves at job fairs, a point that our H.R. person made repeatedly to us in private. I probably talked to 30 or so students, and I could count on the fingers of one hand how many of them approached the conversation in a professional way. It wasn't unusual for students to walk up to us, and the first words out their mouths would be, "So, what kind of jobs do you have for me?" No introduction, nothing. Just a "whatcha got?" mentality.
  • Resumes are weak. Students, by and large, aren't really taking advantage of any opportunities to gain professional experience during their college years -- or if they are, they're not translating that experience into something clear on their CV.
  • We're seeing a real dilemma with hiring new grads. The ones who understand the theory of statistics -- usually, the pure math/stats majors -- understand the how/why of different procedures, but lack the skills to apply that theory to real-world problems (at least in health care). Yet the ones leaving applied programs understand the industry well enough but don't really grasp the how/why of the statistical tests they're using. It's as if they know that a chi-squared test does X, so they look for a chance to do X even if the underlying data don't really support the procedure.

This feels like a conundrum. It's as if there's a gulf between theory and application, and through all of it, we're not really setting these kids up for success after they get one or more sheepskins.

Have any of you who've dealt with new grads seen similar problems? If so, do you have any potential solutions?

I used to think that a solid internship would help, but we've heard horror stories of stats students getting placed into programs where they're basically doing report-writing or data collection. Not meaningful stats, under the watchful eye of a seasoned statistician.

If colleges are failing our new grads, can the ASA step in with some sort of ongoing prof-development program that someone in a new-grad situation would find worthwhile?

2 comments
108 views

Permalink

Tag

Comments

02-06-2014 17:13

With regards to Jason's last question, the ASA is interested in learning about the professional development needs of statisticians that are not being met in other ways. The association has also recently developed a set of recommendations for graduate programs that address some of the issues that Jason raises. See http://stattrak.amstat.org/2013/07/01/preparing-masters/. I hope we'll see these recommendations considered and acted upon by statistics graduate programs soon.

01-02-2014 15:58

I agree with everything Jason says. I've spent a lot of time recruiting for our statistics team over the last few years. I've lowered my expectations as a result. Rather than trying to find someone who will be a perfect fit for everything we are looking for, I go for the most important pieces - foundation of statistics knowledge, communication skills, able to learn quickly, works well with other people. I figure if we can get that, then we can teach them the specific things they will need to be successful at the job.
As a result, it may take a year or two to get a new graduate to where they are competent in doing what we want them to do. I wish there was a way to shorten this time period but I'm less than optimistic about it. One fundamental challenge is that the diversity of work that statisticians do is extremely diverse, both in the tools and applications. It is hard to create a degree program that will cover that large diversity, without increasing the time folks are in the university programs. And the cost of college is making it less likely that folks will want to take even more time for a degree. Not to mention that they may be able learn faster what they need to learn by being on the job, not in class.