What do Statisticians actually do?

[MrScotchpie]

Hello,

I'm currently studing with the Royal Statistical Society as part of a mid-life career change.

I'm really enjoying statistics but I've come to wonder what statisticians actually do. I've begun reading reports that include analysis but I haven't seen any authored by a statistician. They are written by marine biologists, environmentalists or medical doctors etc. usually working as professors with some university.

So if these scientists do their own stats, whats left for a statistician to do?

Andrew

[bugman]

Hi MrScothpie

Are you in London? You might know a friend of mine at the society.

Its an interesting question. Statisticians consult, teach, write, collaborate and research. Some, especially in Government may lead teams in certain departments.

In biology/ecology alot do collaborate with researchers or advise. Statistical methods in ecology have advanced tremendously and alot of biologists/ecologists do develop very good statistical skills. I my area many statisticians have co-authored studies where new techniques have been developed or the study design is so complex that a consultant is hired to advise.


P

[duskstar]

I'm a Medical Statistician, and I can tell you, I do all my own analyses! If you are interested in Medical Statistics as a career I am more than happy to answer questions about that and show you some bits of work I have done so you can get a flavour for it.

The RSS courses have always looked great - I am currently (slowly) getting my CStat accreditation with them

[MrScotchpie]

Thanks for the relies.

@Bugman, your answer clears up a lot of confusion, thanks.

@duckstar, what a coincidence, I work as a data modeller/analyst for the NHS. I was frustrated with the low level of statistical knowledge among my colleagues that I decided to teach myelf via the RSS using the distance course provided by the University of Southampton. My background is in quality management and I hold the CQI certification as well as a level 4 (HNC/HND equiv) in Decision Analysis, so I'm no stranger to stats but wanted a "pure" stats qualification. I'm also hoping to reach the lofty heights of CStat accrediation - I'm currently an elected fellow.

From my experience, NHS analysis is all done in Microsoft Excel, but I am attempting to move away by trying my hand at learning R and Python. My current work involves Stroke/TIA and cardiovascular disease and my main duties include hypothesis testing for significant movements in mortality rates, incidence etc.

Finally I've enrolled on the one month free "Computing for Data Analysis" course provided by the John Hopkins University's Bloomberg School of Public Health, in the USA. As I said its free, online and teaches R programming. To succeed you need to complete two programming assignments. I'm then thinking of following that up with the three month course "Data Analysis" by the same university, which teaches regression analysis, time series analysis etc all in R.

[Jake]

Did you see this free, online, 7-week course beginning on Sep. 24th at Coursera?
https://www.coursera.org/course/biostats

[noetsi]

A high percentage of them stay in universities (which means they write on rarifed topics about a hundred people in the US read). But there is actually a signficant need for them in the public and private sector, which is likely to grow. For example I run statistics (I am not a statistician in the sense you mean) as a regular part of my job to help us analyze our performance and predict future events. Aspects like Six Sigma is making statistics more important than it has been in the past.

We use SAS and SPSS primarily, but we are moving to R. Well I am as the only statistical person in my agency.

[Jacov]

This is actually a good question.

You have to understand that most statisticians / mathematicians do something very different then they have learned. For example, some friends of mine with a PhD in stats or math are now working in banks, as software developer, in the insurance industry… . I would guess that as educated stats/math guy only 20% work somehow in their field where they get educated. Most companies do not hire people because they are experts in lie group theory or in some fancy time series stuff. Most stats / math guys are hired because they are able to think in a logical framework. Also many pure statisticians told me that it was extreme difficult to get a job in a statistical area because often statistician with an extreme good knowledge in biology / finance … are sought.

But this is no exception, I know also doctors you work as medical software consultants. Actually, the only person I know which does all day long regression analysis is an educated pharmacist.

What I want to say is that if you leave the university your working field could become very different from what you learned. Therefore, your question is extremely difficult to answer.

[MrScotchpie]

@Jake, No I hadn't seen that course and it seems to fit well with my current work. Thanks for the link.

[noetsi]

What I want to say is that if you leave the university your working field could become very different from what you learned

This is true of every academic field. It is particularly true in the private sector where they care about concrete issues that can be explained in 5 to 10 minutes max to non-statisticans. I had a conversation once with Vice President with a PHD in economics (a very rare individual to find). I had spent a great deal of time creating a "better" (in terms of providing correct information) metric. He said to throw it out and use a less "correct" one because it was simply too complicated for its (high level) audience. And I don't have the involved knowledge of those here so it was not complex to any one with strong statistical knowledge.

Think simple and practical if you use statistics in the private sector.

[Dason]

We fight terrorism of course.

[SmoothJohn]

I'm a philosopher who does statistics for (part of) my living. It's a funny world.

[Dason]

SmoothJohn go get 5 more posts quick! I'll give you a treat if you do.

[trinker]

Essentially a statisticians job is to make a mountain out of a mole hill Tongue in cheek.

Send all hate mail to Dason K, U of Iowa, Cornfield, IW 55555

[Dason]

Where do we send the hate mail for associating Dason with the University of Iowa?

[SmoothJohn]

I am intrigued, Dason.