https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim Birds-of-a-Feather Friday 20MAY 12:00-12:30pm (Double-click anywhere to open up your own window to leave a comment.) en-us 2016-05-19 00:51:14 UTC 2025-10-05 07:58:57 UTC hello@padlet.com https://padletuploads.blob.core.windows.net/aws/92578/8031a24d6035c32abd172ec947465c9ecd93ddcc/a8cadc67104e3ae64b50fc5214fd3b18.png https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111660940 Please type questions or independent comments here, along the top.]]> 2016-05-19 13:35:49 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111660940 https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111661086 Please reply to questions or comments below them, to make vertical conversations.]]> 2016-05-19 13:36:27 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111661086 https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111888553 Faculty development is critical for stat ed in general and simulated-based inference in particular.  MOOC's and other online learning opportunities seem well placed to address some of this, but more is needed.  Any guidance about how to help this happen? (P.S. eCOTS is a key component!)]]> 2016-05-20 16:04:31 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111888553 https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111888691 A good deal of time is spent in the traditional course talking about the assumptions for each analysis and thinking about whether they hold. I think that's always good to do.  In simulation- based courses do you lose much of that?]]> 2016-05-20 16:05:12 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111888691 https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111888779 Do we have a clear consensus about what faculty need to know?  In particular, how much must they know beyond what is covered in the courses they teach?

]]> 2016-05-20 16:05:45 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111888779 https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111889050 The 1st time I saw bootstrap was in graduate school, in an applied  Statistics course.  Do you think freshman students' are ready?

]]> 2016-05-20 16:07:27 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111889050 https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111889391 I think the situation is better in an SBI course because (1) students can handle more situations and (2) they have better tools for checking assumptions.]]> 2016-05-20 16:09:06 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111889391 https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111889464 This is a great question, Randy, and applicable across all teaching of statistics and not just SBI. The closest thing we have to research in this area are a few people working on TA knowledge for teaching statistics (Jennifer Noll, Jenny Green, Erin Blankenship come to mind and I am starting to work in that area as well). In any case, this is a great area of young statistics education researchers.

]]> 2016-05-20 16:09:26 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111889464 https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111889529 You still need to worry about conditions - especially when moving to the traditional approximations.  It's nice that you can really focus on those points at that point.]]> 2016-05-20 16:09:51 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111889529 https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111890039 Are they ready for normal theory?  In each case, you need to teach it at the level of the students you have.  One advantage of the SBI approach is that it feels more grounded.  (The formulas in a traditional course are hard to justify to students in an Intro Stats course.)]]> 2016-05-20 16:12:15 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111890039 https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111890301 SBI is included in the Common Core Standards for Mathematics at the high school level. If the Common Core is taught as written, college freshman will have seen SBI prior to matriculation.]]> 2016-05-20 16:13:59 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111890301 https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111890311 2016-05-20 16:14:02 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111890311 https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111890344 Randy's point 1 works against what I was talking about (having students view a technique as providing a "silver bullet").  His point 2 is key.]]> 2016-05-20 16:14:16 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111890344 https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111890522 Regarding point 1: It is frustrating in the standard approach that students have so little else they can do when things fail.  So being able to handle more situations is a plus.  But as Robin says, it does not remove the need to worry about conditions.  And as you say, point 2 makes that better.

]]> 2016-05-20 16:15:25 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111890522 https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111890566 Also there are a lot of researchers (e.g., in the SRTL group) that have seen pre-college students learning SBI (or at the very least motivating it with explorations of sampling variability).  See the work of Dani Ben-Zvi, Cliff Konold....  I think also Maxine Phannkuch and Chris Wild have done informal inference using simulation with middle school students....
CORRECTION!  Chris says they only really did SBI with high-school students but the motivation (E.g. wiggling box plots) was with middle school students.]]> 2016-05-20 16:15:41 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111890566 https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111890787 One issue is that most teachers were not taught statistics from an SBI perspective]]> 2016-05-20 16:16:20 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111890787 https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111890899 Agreed]]> 2016-05-20 16:17:03 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111890899 https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111891125 2016-05-20 16:18:13 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111891125 https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111892244 Good place to go for more resources/discussion on teaching with SBI
causeweb.org/sbi]]> 2016-05-20 16:24:39 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111892244 https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111892498 This means we have to be preparing students to teach SBI as part of the common core]]> 2016-05-20 16:26:16 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111892498 https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111892594 And prepared for students for whom SBI is not new/fresh when they get to college courses because they will have seen them in high school.]]> 2016-05-20 16:26:54 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111892594 https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111892622 With good software supporting it, intro students appear to pick this up with little difficulty.]]> 2016-05-20 16:27:05 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111892622 https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111918906 At Macalester College, we have a data-science minor. The introductory course attempts to orient students to what data science is. The course uses a book, "Data Computing" (see http://data-computing.org). 

Things to note:
* the book uses a very compact set of R commands, in the spirt of Randy Pruim's "Less volume, more creativity" approach from his Tuesday eCOTS session.
* the book is intended for a short course, about half a full course. For a full course, this can be supplemented with either more projects (data cleaning? data scraping?) or topics such as statistical modeling or machine learning.

This note isn't entirely a shameless plug. I have some copies that I'm happy to loan out to people, but first check the web site to see if you're really interested. Write to kaplan@macalester.edu to request a copy.]]> 2016-05-20 18:55:09 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111918906 https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111956765 2016-05-21 13:14:37 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111956765 https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111970187 One nice aspect of conditions with SBI is you can SEE the distribution of the statistic (e.g. bootstrap distribution)  and directly assess whether it is skewed.  This is much more concrete than students trying to guess at whether some combination of n and skewness in the data is enough for the CLT to work it's magic.

Also, with randomization tests the only real condition is randomization!  No sample size or distributional requirements whatsoever! :)]]> 2016-05-21 18:55:54 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/111970187 https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/112014586 I don't think a whole MOOC is needed (if they already understand inference).  CAUSE has some good webinars on both randomization tests and bootstrapping.]]> 2016-05-22 17:38:00 UTC https://padlet.com/ekendriss/eCOTS16_BOF_TeachwithSim/wish/112014586