Self-Evaluation and Lessons Learned (Introduction to Statistical Computing)
Attention conservation notice: Academic
statistico-algorithmic navel-gazing.
With the grading done, but grades not yet posted while we wait for the
students to fill out faculty evaluations, it's time to reflect on the class
just finished. (Since this is
the third time I've done a post
like this, I guess it's now one of my traditions.)
Overall, it went a lot better than my worst fears, especially considering
this was the first time the class was offered. There was a lot of attrition
initially, both from students who had taken a lot of programming, and from
students who had done no programming at all. (I was truly surprised by how
many students had never used a command-line before.) The ones who stuck around
all (I think) learned a lot --- more for those who knew less about programming
to start with, naturally. Most of the credit for this goes to
Vince, naturally.
Some stuff didn't work well:
- The labs were too hard to finish in 50 minutes. (Every student who
mentioned the labs in their feedback, and that was most of them, complained
that they were too short, and that there were too few TAs.) Either the
problems need to be made much easier, or we need much more lab time, or we need
to ditch labs. (But it would be good to give them immediate feedback
on programming...) I am not sure what the right thing to do is.
- The in-class midterm. This did not probe the
student's skills as well as I'd hoped, and the very low scores seem to have
depressed morale. (It got curved, of course, so it didn't end up hurting
anyone, but still.) Next time, either a take-home midterm, or eliminating
the midterm altogether in favor of more weight, and time, on the project.
- The final projects need more time, and more intermediate feedback for
mid-course corrections.
- Writing problem sets the weekend before they were assigned. (I don't
think it will surprise any of The Kids to learn I was doing this, or that Vince
was better organized.)
- If the word "hate" was uttered each nanosecond of the hours I spent
wrestling with Blackboard, it would not equal one one-billionth of the hate I
feel for that software and its
designers at this instant. Unfortunately I don't have a better solution
which (i) lets students submit their work electronically, (ii) lets the
graders share the work, and (iii) provides a shared gradebook.
Stuff that worked well:
- Most of the homework assignments, despite the visible seams.
Specifically, writing the assignments as (very nearly) a series of tests seems
to have helped, and should be pushed further. (I got this idea
from Bill Tozier, though he may
not recall it.)
- Teaching testing and top-down design. (Grading should enforce this more
in the future.)
- The data-wrangling topics were a big hit. (Again, all to Vince's credit.)
- Giving a group project final instead of an in-class or even a take-home
exam. (Results speak for themselves.)
Stuff I'd try to do next time:
- Provide more hints about looking stuff up in The R Cookbook.
- Require the use of a sensible text-editor from the beginning (and maybe
have the first lab be mostly about that, plus introducing the command line).
RStudio would probably work, though to be
effective I'd have to switch to it myself, and away
from R.app + Emacs.
- Enforce style, naming and commenting conventions even more rigidly than
now (especially commenting).
- Schedule project presentations during the final exam period, so they don't
eat into lecture slots. Between that, and not having to kill lecture slots for
the midterm exam, the pre-exam review and the post-exam inquest, it should be
possible to add back in optimization, more about simulation/Monte Carlo, and
even more data manipulation.
- Clarify expectations at the beginning: students will have to use
statistics they already know from the pre-req classes, and to learn new
statistics. (It's not as though there aren't plenty of statistics-free
programming classes for them to take...)
Over-all assessment: B; promising, but with clear areas for definite
improvement.
Obligatory disclaimer: Don't blame Vince, or anyone else,
for what I say here.
Introduction to Statistical Computing
Posted by crshalizi at December 20, 2011 09:35 | permanent link
