David Wooff writes, "If your students can't learn a tool such as R, it's
a pretty good indication that they can't learn statistics", responding
to a question on software to draw graphs. There are so many issues
rolled up in this question and responses!
1) Introductory statistics courses for other subject-majors have always
been unpopular. In the past they were devilled by students getting
bogged down in the arithmetic, now they get bogged down in the IT. It
is very hard to take in statistical concepts when you are completely new
to the tools. Hence most students maintain or adopt the cookbook
approach to statistics.
2) R is a completely general computing language. While it can be
shielded with a higher-level interface, it retains the flexibility to
bite back. Users need to understand computing concepts (structures,
loops, data representations) and beware unintended consequences. Hence
in my opinion it is a very poor choice for teaching statistics from
scratch. For similar reasons, it is a dangerous tool for amateur data
analysis - treatment of missing values being one elephant trap.
3) Since students hope to go on to careers that will depend on data
analysis over several decades, it is a very strange philosophy to start
by inculcating the idea that their critical tool in this area should be
"free" as its primary criterion. I don't know how we get it through to
top managers, politicians and academics that having paid for the
hardware and a large licence fee (non-transferable) to Microsoft,
further software should somehow not incur any cost - apart from hidden
costs of wasted time in unrecognised training and support. Don't get me
wrong, much of the best software is "free", but then why don't
universities promote Linux?
4) Most senior managers assume Excel is "free"! It has its proponents,
and can draw graphs. It also has drawbacks (well documented) and a
great set of elephant traps for the unwary. For example, it probably
requires some data re-arrangement to "plot Y against X, then plot again
with apparent outliers excluded". I have plenty of reservations about
its graph drawing, but "it's there".
5) I take David's point, in that one of the reasons behind so much
misuse of statistics is that students take "Stats 101" once in their
life, dislike it but remember a few basic recipes, and pass that
attitude down to future generations. Perhaps Margaret is looking for
something that will use more graphical examples to convey the ideas of
statistical thinking to students, so they start to see some general
logic in analysis rather than get confused and annoyed by the detailed
steps. Would it work better to teach "statistics" using computer
output prepared earlier or run by the teacher, then have a follow-on
course for those who want to learn to use the tools?
Allan
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|