What's higher level statistics like?
April 6, 2010 12:43 AM   Subscribe

Math-related career advice requested: is work in higher level statistics compatible with an imaginative-type thinker or would it be total hell?

I majored in physics in college, am currently a teacher, and am playing with the idea of applying to a masters in statistics program, and then perhaps working in the private sector or for the government.

Can anyone tell me what higher level statistics is like? I'm trying to figure out if I would enjoy doing it as a career. (I've read David Hand's "Statistics: A Very Short Introduction" to catch a glimpse of how statisticians work.)

Some background:

In physics, I excelled in courses that demanded some visual sense, like E&M, Stokes and Gauss's law, complex analysis, etc. On the other hand when I took a year course in calculus-based probability I found it challenging (although I was able B to B+ marks)--I found it very logical and often non-intuitive.

However, I enjoyed imagining the data distributions and real-life scenarios posed. The class also made me appreciate the power of statistics. I often think to myself about all the amazing things I could possibly solve if I only had advanced knowledge of the subject.

I'm more of an intuitive, big picture, visual thinker. I enjoy analyzing systems, understanding conceptually how things "fit" together (not as an engineer would, but as a writer would his or her work), and sometimes not so good with details. Recently I took a Myers-Briggs-type assessment and came out INTP.

To summarize, is work in higher level statistics compatible with an imaginative-type thinker or would it be total hell?

If hell, what other math-y fields do you think I'd be more comfortable working in (excluding physics)?

Thanks for your help!
posted by pinside to Science & Nature (8 answers total) 20 users marked this as a favorite
 
Do it. Probability and Statistics are lovely subjects. An introductory statistics course normally seems kinda dry & boring for math and physics majors who'll be more into romantic stuff like functional analysis, differential topology, algebraic geometry, optics, relativity, quantum mechanics, etc. But statistics gets way more exciting once you start talking about experiment design.
posted by jeffburdges at 1:19 AM on April 6, 2010


Another option would be acturial science.
posted by sanskrtam at 1:57 AM on April 6, 2010 [1 favorite]


I don't know about work, but higher level statistics-based research definitely is a joy for imaginative people. It's a lot of grunt work, and to do well you need to be quite organized/methodical, but it's pretty fun to play around with variables and different hypotheses. If you like economics, you could also go in an econometrics direction.

One thing I always hated, though, was reading awesome papers and finding the one tiny forgotten statistical detail that derailed the entire project. You might enjoy this book, written by my favorite professor Ray Fair (who wrote a model that correctly predicts every presidential election). His website (especially the econ 483a part) has lots of fun links and things to play with.
posted by acidic at 2:53 AM on April 6, 2010 [3 favorites]


Seconding research design / biostatistics, maybe even an MPH/MSPH. That will keep your coursework well-centered on applied stats. For extra fun, look for bayesian-heavy courses.
posted by The White Hat at 5:58 AM on April 6, 2010


Sent you a MeFi Mail.
posted by dsword at 7:55 AM on April 6, 2010


Modern statistics is a broad field... There are some applications that are very "details-oriented" (eg sampling design, sometimes) and sound like they might be less appealing to you. OTOH, some areas of statistics are extremely visual: exploratory analysis with multivariate data, or various machine-learning applications.

You mention that you're a big picture thinker and like analyzing systems... You might have most luck being able to do big picture thinking if you pick a certain sub-field (like physics, ecology, public health, whatever) and concentrate on developing yourself as someone who works on statistical problems in that area. Possibly even consider doing a dual MS program in both statistics and some other science. You'll have less opportunities for big picture thinking if you are coming in as a statistical consultant on a topic that you don't know as much about.

Also, there are some great blogs out there about (visual) data analysis and statistics:
http://flowingdata.com/
http://infosthetics.com/
posted by JumpW at 12:29 PM on April 6, 2010


Another blog to check out, by a Columbia prof, focusing on social science topics applications of statistics. It might be interesting to you precisely because the author clearly has wide-ranging and creative interests, and uses statistics as a window into them.
Statistical modeling, causal inference, and social science
posted by slipperynirvana at 4:44 PM on April 6, 2010


Thanks everyone for the helpful guidance. I'll be looking more into this and checking out some of the leads and links provided. I'll also periodically return to this thread to see if more people have added advice.

I subscribe to FlowingData's feeds and have perused some of Tufte's books. I like the idea of using statistics in an almost journalistic way--investigation, illumination, and finally through visualization, communication. On the other hand I also find exciting the possibility of working in a field like statistical genetics. So I'm thinking one could do a lot with a deep knowledge of statistics.

I guess I play with the fantasy of doing the type of work described above but don't know if the day-to-day work to accomplish that fits my way of thinking or personality.

Thanks for the encouragement!
posted by pinside at 12:48 AM on April 7, 2010


« Older How to save email files to SD on Mytouch 3G phone?   |   Lights off. Heater on. What? Newer »
This thread is closed to new comments.