[Turnbull Zemi] Tomorrow's Zemi
Stephen J. Turnbull
turnbull.stephen.fw at u.tsukuba.ac.jp
Thu Sep 19 00:12:01 JST 2019
Hi everyone,
I'm sorry for writing only in English, but I'm very tired and left
this until late.
We are now preparing the M2 students for their midterm presentations.
Those students are 王 曉飛, 張 子雲, 蒋 穎, 鈴木 康介 in an order
to be announced tomorrow. Discussants will also be assigned
tomorrow. Discussants have no preparation to do, and do not need to
receive presentation materials. See "Discussants" below.
Presenters
----------
The presenters should present as much of their research as possible.
Do not worry about a time constraint, you may use up to 30 minutes,
but you shouldn't *try* to use that much time: between 12 and 20
minutes is best. The following areas should be covered:
- These slides are the basis for your midterm presentation. Make
every revision as perfect as possible in form, because you may not
have time to change it later.
- Title and self-introduction
- Background and motivation of the research.
Why did you become interested in this theme, and any interesting
aspects for science, society, or policy (*except* previous
research).
- Literature survey.
This includes previous scientific papers in your field. Usually
there is a seminal ("seed") work which is famous, and one or more
paths made of several papers each leading directly to your thesis.
How you describe these paths is up to you, but usually it's
"depth-first", that is, you start at the beginning and cover each
paper in one thread until you reach your paper's original idea, then
return to a branch point and cover another thread.
- Model.
This describes the variables and relationships among those variables
that you are studying. Usually it is a modification of a model
presented in an earlier paper.
- Data. [skip if theory paper]
This describes the source and method of measurement of each
variable. If you are using data published by someone else, source
is easy but you must be able to describe the method for gathering
data as if it were your own.
- Statistical methodology. [skip if theory paper]
This describes the statistical methods you are using, including
parameter estimation and hypothesis tests. If you are using a
questionnaire survey you should be doing reliability tests, factor
analysis. Hypothesis tests should include power (probability of
Type II error, accepting the null hypothesis when it is false). The
computer programs you use are *not* methodology. You *should* cite
them, however.
- Estimation and test results. [skip if theory paper]
Describes the results you achieve with your statistics.
- Calculations and proofs. [skip if empirical paper]
Describes your analytical methods for calculations and proofs and
any theorems or results.
- Implications of your results for scientific progress or policy.
- Suggestions for further research.
- Reference list.
- Appendices.
These include any instruments (eg, questionnaires) you have
constructed yourself, full data sets if practical else reference to
online sources, detailed proofs or calculations, additional graphs
and tables, and any other materials that don't belong in a
presentation but might be useful in answering questions.
These areas do not correspond to chapters in your thesis, and not
always in exactly this order. However, this is a fairly conventional
order.
Discussants
-----------
Discussants should focus on the presentation materials, not on the
content. If you notice anything in the research itself you want to
ask about, that's fine, but your main job is to check format.
- Are page numbers visible on every slide?
- Are titles informative? For example, better than "Implications" is
something like "Effects on the Trade Balance". Also, in a full
presentation some sections will require several slides. Each slide
should have a unique title.
- Are fonts readable? Many faculty members are older and don't have
great eyesight. If they can't read it from the back of our lab, it
*does not belong on a slide*. Exceptions include the reference list
(which nobody except me reads until your final version) and large
tables.
- Does every table and graph have a caption?
- Is all text on graphs readable? Do the colors help you to
understand? Are all needed labels present (variables on axes, tic
marks and values for graphs of data, functions that are visualized
as curves, any curves or lines that divide a graph into regions,
etc?
- Can you understand what the graph means without the presentation?
Check this before the speaker explains!
- Citations should be in the form "Author last name[s] (year)". Do
not include titles and publication data other than the publication
year (that's what the reference list is for). When you cite a
paper, explain why you need to cite it. For example, "Nick and
Percy (2019) first applied the Furgleblatz everted regression method
to triply-redundant Poirotian data. Their methodology is therefore
most appropriate for this research based on triply-redundant
Poirotian data." *Never* use numerical references such as "[1]" or
"Percy [1]" in a presentation. *Do not* make the audience flip back
and forth between the content and the references.
- Does the time used for each slide seem way too little, or way too
much? If it's too little, should the slide be deleted, or does it
need more discussion? If it's too much, should the slide be split
into more than one, or should the discussion be trimmed?
--
Associate Professor Division of Policy and Planning Science
http://turnbull.sk.tsukuba.ac.jp/ Faculty of Systems and Information
Email: turnbull at sk.tsukuba.ac.jp University of Tsukuba
Tel: 029-853-5175 Tennodai 1-1-1, Tsukuba 305-8573 JAPAN
More information about the G-Zemi
mailing list