Date for
Final Presentations
Here’s the
list of people who have signed up thus far:
Mon Nov. 27th:
Octavia
Eleni
Tyler
Mohammad
Victor
Mon Dec 4th
Shamina
Charles
Ana
Taylor
Leah
David
Hasan
Brandan
Karam
Today is the
final group project presentation on Multivariate statistics.
1) Following
that, we will not meet for the next two weeks, but we will meet again on Mon.
Nov 27th and Mon. Dec. 4th
where you
will each have a 15 minute timeslot to present your individual project (see
details below).
We’ll split
the class into two, with one half the class 7 people presenting on each day.
The talks
should be as if it was a presentation at a conference. So do indeed set up when
the biological
Questions/hypotheses
issues are. But unlike a conference, you’ll mostly go somewhat “overboard” in
Presenting the statistical analysis of your data.
So, I’ll
need “volunteers” for each of the two dates.
2) Your
project write-ups will also be setup as if they are a scientific publication,
but again you’ll go heavily overboard on the statistical side.
Introduction: So, do begin with the biology/intro
– perhaps 1 page or less.
Methods particularly as they pertain to the
experimental design and statistical analyses.
So, for example, we don’t need to know what
buffers were used and all that sort of thing. Nor how gels were made etc. Or how you caught salamanders etc.
Results: The results section will likely be
the longest part especially because you’ll have in it lots of graphs of the
data, of appropriate kinds
(e.g. histograms, scatterplots and the like ) with
descriptions of what they show.
(e.g. a histogram of the distribution of clutch sizes shows a
skewed distribution with most nests containing 1 egg).
In terms of
the analyses, begin with the simple – which would include descriptive
statistics and graphs, and then move
To the more
complex (e.g. if relevant you might end with your multiple regression, or 2 way
ANOVA or whatever your project is mostly about
Also ensure
to include stats where you test whether or not you met the assumptions of the
tests you carry out.
Discussion/Conclusions
This
sections will be perhaps 1 maybe 2 pages, highlighting the findings and
focussing on the stats, but referring back to the biological question, in the
end too.
So, e.g. we
found it necessary to assess nestling survivorship using non-parametric mann-whitney as the date were
markedly non-normal in their distribution. The MannWhitney
showed
That
survivorship increased under the forest canopy compared with the oldfield - or whatever).
Grading scheme for course.
Group
presentation 30%
Individual
presentation 30%
Project writeup
40%
Project due
on Dec 8th!
Below are
links to documents provided by the various groups presenting in the course.
Anova
documents click below
Multivariate (note, that variables are cells, and the genes are subjects. It is also likely that the PCA loadings were used in the cluster analyses).