Ista,
On the one hand I'd like it to be as flexible as possible so the
students could really come up with whatever they like. On the other
hand, restricting their choices probably would make it easier to do the
backend. The goal would be to get them to realize that the apparatus of
hypothesis testing (when done via simulation/randomization techniques)
doesn't depend on what the statistic is. The flow of steps is the same
whether the statistic is a mean, variance, or their own kooky thing.
Obviously this isn't the end of the story - what the statistic is
actually describing is also a crucial component to interpreting the
results of a hypothesis test, but I think it teaches an important
pedagogical point about where statistics come from and that if they find
themselves in a situation in the future where they need to make up their
own, then that is perfectly okay.
So they could make up ones like: arctan( (max({x})^2)/ (min({x})^2)
)-3), max({x})-min({x}), sum from i to n of (x_i - 25th%ile({x}) )^3)
[that might be tricky to write in a standard equation editor]
none of these is hard to write in R, but translating from an equation
editor might be.
Perhaps the best solution is (as Alan suggests below) to write a Shiny
function builder myself so that I can control the whole process and make
sure that they can't enter anything that would break the backend. Or to
have them learn the rudiments of writing equations in R so that it
bypasses the whole process.
Thanks,
Scott
On 11/19/14 1:52 PM, Ista Zahn wrote:
Hi Scott,
Can you give a couple of examples of the equations you have in mind
along with how those should be translated to R?
Thanks,
Ista
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
