8
$\begingroup$

I'm moving to a new university for the fall semester, and I'll be teaching a statistics class for the first time. I'm familiar enough with doing statistics (my dissertation in math ed was a mixed-methods study with some quantitative components). However, I've never taught it before, so I don't yet have that level of knowledge about key statistics concepts and/or the "right" ways to think about them.

To give you a for-example: When I teach calculus, I know that I want people to leave understanding (among other things) that the subject is really mostly about creating and refining approximations (the limit process; linear approximation; the Riemann sum; etc. etc.).

So, in an effort to bootstrap my understanding a little bit, I thought I'd ask here: What are the big ways of thinking that are really useful for understanding and properly using statistics?

If it's helpful to know, our book is OpenIntro's Intro Stats with Randomization and Simulation.

Thanks in advance!

(FYI, I've crossposted this to /r/math.)

Improve this question
$\endgroup$
10
  • $\begingroup$ I like to approach statistics as the answer to Hume's Problem of Induction. $\endgroup$
    – Craig Feinstein
    Commented Jul 5, 2018 at 16:28
  • $\begingroup$ @CraigFeinstein I'm intrigued! Say more! $\endgroup$
    – Spencer Bagley
    Commented Jul 5, 2018 at 16:46
  • $\begingroup$ see my question and also my answer: philosophy.stackexchange.com/questions/39545/… $\endgroup$
    – Craig Feinstein
    Commented Jul 5, 2018 at 17:17
  • 4
    $\begingroup$ If you don't get enough useful answers/suggestions here, perhaps try stats.stackexchange.com $\endgroup$
    – Yemon Choi
    Commented Jul 5, 2018 at 17:44
  • $\begingroup$ Bayes versus frequentist. (And Bayes is the proper way... when conditions are met.) $\endgroup$
    – David G. Stork
    Commented Jul 5, 2018 at 17:51

2 Answers 2

Reset to default
3
$\begingroup$

If a student walks away understanding the distinction between an estimand, an estimator and an estimate, along with the idea of a sampling distribution, you will have done an above average job. Equipped with these concepts, it should be clear that one must evaluate estimators based on how well they perform on average, in some relevant sense (such as coverage or mean squared estimation error).

On a more personal note, I favor de-emphasizing testing and especially cook book recipes for statistical tests --- it encourages missing the forest for the trees and arguably is not especially relevant in many applications.

Improve this answer
$\endgroup$
2
$\begingroup$

The American Statistical Association has resources for statistics education, available here:

http://www.amstat.org/asa/education/home.aspx

...including, in particular, resources for undergraduate education.

Perhaps the "Guidelines for Assessment and Instruction in Statistics Education (GAISE) College Report" (available under the "Guidelines and Reports" heading) might be helpful. Oh, here is a direct link, why not:

http://www.amstat.org/asa/files/pdfs/GAISE/GaiseCollege_Full.pdf

This report seems relevant for your question because it lists 9 "big" goals for students. From there you can follow up to get more detail about those goals; recommendations for how to help students work toward the goals; pointers to further resources; etc. Good luck!

Improve this answer
$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .