Home » Columns, Featured, Master's Notebook

In the Hot Seat: Two Experienced Consultants Discuss Hiring Statisticians

1 July 2010 7,781 views 7 Comments

    Schweitzer: “For example, with regression modeling, I want to know if they did any model validation. Did they do any diagnostics with their model? You know, I have interviewed people who have run a regression analysis, but never did any diagnostics. To me, that really says that they didn’t really know that much about what they needed to do.”

Data Analysis

Both PwC and E&Y’s practices rely on working with large data sets. As such, SAS or other database programming skills are important.

    Schweitzer: “We have hired people who don’t have any SAS experience. [In those cases,] we like to see that they have at least worked with large data through some other package. [When reviewing their skills,] I ask about the size of the data sets they have worked with. If they are still in school, chances are they have only worked with fake data, which you know is relatively nice and clean data sets. A big piece of the job is getting your data ready for the analysis, and a less amount of time is spent actually doing your analysis.”

    Batcher: “I also look for internships and work experience where they’ve learned SAS or other computing packages. Most everyone puts down that they know SAS, but sometimes that means they have taken a class, the instructor has given them code, and they altered the code slightly and ran it. But if they’ve used SAS and actual[ly] written the code themselves—in a real environment —[it] is a plus.”

Both Batcher and Schweitzer mentioned that statistical programming skills are less important with management positions.

    Schweitzer: “As you move up, [database programming] largely becomes supervisory. For example, I do not have SAS installed on my laptop. I still need to understand the day-to-day analysis, but working with the data is something that I would delegate.”

Communicating Statistics

Batcher and Schweitzer both find communications skills to be a make-or-break characteristic in the consulting world.

    Batcher: “Our clients need to understand [the candidate]. Clients can become very frustrated if they cannot understand the statisticians. When we communicate statistics, there are two things I want to make sure doesn’t happen. One is showing off and using too much statistical jargon to our clients. The second thing is adding too many details [when describing an analysis many statisticians say,] ‘Well, first I did this, and then I did this, and then …’ I prefer a two-sentence explanation or overview in language that would be clear to our clients. In fact, for many of our clients, this is sort of their first encounter with statisticians. So, candidates’ explanations should be enough that they get an intuitive grasp, but we don’t want to drag our clients through the weeds. [Statisticians] love to throw out ‘this or that, but that would then violate the statistical assumption that blah blah blah,’ which is totally meaningless to the client. So, it seems like we, as statisticians, speak this mystical language and therefore you, as a client, are dependent on me. And that’s what I consider to be showing off. You can explain the details, but you need to say it in a different way, such as, ‘Well, we can take a couple different approaches and they each have their own set of mathematical assumptions … we’ll find the best approach which will meet all of the assumptions for this project.’”

    Schweitzer: “I’m looking to see that they understand the results of the analysis and understand the implications of the results. If they worked on an interesting project, I’ll always ask what the outcome was. I sometimes think that statisticians stop and don’t really take that next leap to show what the analysis means, which is important for consulting. It’s not just what test did you apply, but also what was the outcome and what did it mean for your employer or client, and what actions were taken as a result of your analysis?”

    Batcher: “Sometimes I might just ask them to explain [a statistical concept like what is a standard deviation]. Sometimes I’m surprised; I’m not sure if it’s the nervousness about the situation, but … they stumble around quite a bit. I think that when people get a little flustered, it might be smart for them to take a deep breath and collect their thoughts instead of just jumping right in without thinking the answer through a bit. Now, if they cannot answer the question, it’s not a deal breaker, but if they can do it, it’s a plus. I guess we’ve all been in situations where we botched a simple question. So if everything else is stellar, it won’t kill them [, but pausing before rambling may increase their chances of success].”

Even before the interview, Batcher and Schweitzer both look for evidence of communications skills.

    Batcher: “I look at the cover letter. I look at the résumé on the coursework and sometimes you can tell that they took courses that included writing and that they participated in activities. Also, [work in teaching or] consulting labs are good.”

    Schweitzer: “I like to see evidence that a person can talk to a client. So whether that is having some kind of extracurricular activities or work with the community, you know something that shows that they are not just interested in doing their statistical work, because when you are in consulting, talking to a client is one of the most important things you need to do. Yet, finding someone with those skills can be difficult.”

Moving from Analyst to Manager

Both PwC and E&Y have a solid hierarchy that starts with analysts/associates/staff and moves up similar to the following: senior associate, manager, senior manager, director, and partner. Batcher and Schweitzer’s groups tend to hire primarily associates through managers, since other positions require a large sales portfolio. Tanenbaum, Schweitzer, and Batcher discussed how Schweitzer and Batcher’s search changes based on work experience or level.

    Batcher: “Most of what I’ve been thinking about is right out of school. The candidate is finishing in a month or two. They have a couple internships; they have their transcripts. We have an example of a class project; we have their résumé and cover letter and we are looking for three things: (1) a solid foundation, (2) the ability to communicate, and (3) less about specific knowledge. What I am looking for changes at the management rank. At entry management level, do they know how to help someone when they get stuck? At the next level, I want to know that they know how to guide a bigger project. They don’t have to run giant projects; they may just be managing their own work, but they need to know how to pay attention to deadlines or coordinate with clients and take the lead that the project goals are met, including the timelines.”

Schweitzer’s group hires fewer candidates straight from school.

1 Star2 Stars3 Stars4 Stars5 Stars (4 votes, average: 5.00 out of 5)
Loading...

7 Comments »

  • Frank De Meo said:

    I appreciate Erin putting this article together. There is a lot of master level statisticians in the work force. I agree with Mary’s comment about mathematical statistics. Knowing where we stand in the overall statistical community is encouraging.

    Frank De Meo

  • Paul Teetor said:

    Great article! Thanks for taking the time to articulate the perspective of the hiring manager. That’s very useful.

    I notice that both these manager work for large firms. Do managers at smaller firms, such as a statistics consulting “boutique”, have a different perspective? That would be a great follow-up article, since some of us prefer working for smaller organizations.

  • R said:

    All said and done, these consultants’ world is a very tiny piece of real world. You cannot generalize this in hospital, biomedical, R&D, governmetal or NIH type consulting which requires much more technical expertise than social skills.

  • Clay said:

    Just wanted to say I found this article extremely valuable and insightful! I am graduating with a master’s degree in Statistics this December and already have two interviews lined up. I’m getting nervous about them but reading this article has settled me down. I will definitely bookmark and reference in the coming months.

    Great idea for an article. Keep up the good work.

  • Rian said:

    Just wanted to comment that I thoroughly enjoyed the article.

  • Erin Tanenbaum said:

    Thanks! The article was a lot of fun to research and then write. I worked with both Mary and Laura prior to moving into market research. Even so, I was both surprised and excited by many of there answers. Most of their responses ‘made sense’ but are often not spoken aloud.
    I agree with Paul Teetor – it would be interesting to have a follow-up article either on smaller practices – or maybe even consulting labs. In addition, I’m now curious about other hiring procedures for areas like those that “R” mentioned. Also, I agree with R’s comment that one should not over generalize. Still, some pieces of this article are fundamental: technical training match needed skills? Check. Confidence to answer questions on the spot? Check. etc., etc.

    Thanks, I enjoyed reading your comments.

  • Prabha said:

    Very informative article for all Masters students. Thanks!