What real life statistician's job look like?

Question

I have recently finished statistics course and would like to know if statisticians really do what we covered in the course (usual college level stat course material). The course made me interested in statstician's work. But I have questions, that are not answered in simple "statistician's work description".

Does your job consist more of checking using t- and F-tests, etc.? Or is it closer to academic tyoe like trying to find the way to reduce bias, etc.? For example (dumb example), if you are testing accuracy of predictions are you considering going deep into assumptions made by the predictors (e.g. constant prices,etc.)?

Basically, if you are statistician, could you please share with me (us, at stackexchange) your work routine (I mean things one would not see in typical "statistician's work description")?

Answer 1

Honestly, the actual analyses take very little time. Most of the effort, at least for me, is spent on the following, with my rough estimates:

1. Understanding the question that needs to be answered and the required data $\sim 25\%$. Often, the client needs help going from a "mess" to a well defined problem that can even be answered with data.
2. Collecting, validating, and "cleaning" data $40\%$. Data, especially in my area of environmental risk consulting, is messy: there are errors, "censored" values, inconsistent units and notations, different lab methods, etc. All this needs to be harmonized to allow at least an approximate "apples to apples" comparison of different data.
3. Deciding and performing the analysis $10\%$. Once you know your data types and question, then there are only so many ways to construct a valid analysis.
4. Interpreting and communicating results. $25\%$

In general, expect the 'soft skills" to be used way more often than the number crunching, but you'll always be thinking "numerically" and trying to put numbers in context. Its just hard to test that ability in a classroom...it really comes with experience.