🚀 go-pugleaf

Dear reader, hi there,

I am an probationary teacher in Westfalia {according to the German education
system, the attendence to a university for
ca. 6 years is followed by a 2-year-phase of teaching praxis (called
'Referendariat'), which leads you to the 2nd state's exam}. 
I am about to write my final thesis in a 12th grade course on statistics. 

The topic reads: "Introductionary treatment of assessing statistics by means
of a taste test". I am planning a sequence of 5 or 6 lessons (per 45 mins),
starting with a taste test (choclate I think) and leading to a first contact
with testing theory. 

As approximating tests won't be possible (DeMoivre-Laplace etc. are out of
reach -- we just introduced condional probs), I will choose an exact test.
Besides an binomial test statistics, I consider Fisher's exact test as most
appropriate: the hypergeometric distribution is well known from Lotto, and we
discussed fourfold tables recently (the question could be, then: do women have
a better taste? <the course consists out of 6 girls + 10 boys aged 18>).
Additionally I read an (german) article propposing a rank sum test a lá
Wilcoxon. To me, this seems to be the worst option.

What are your experiences/oppinions on the question of the most appropriate
test? I'd be really very glad about any statements.

Thank you anyway
Argee
        EMAIL:  rgx (at) gmx (dot) de>
===
Note: in case you are using the auto-reply-button: please, omit the
NoSpam-insertion in my reply-to field!