LBSC 878 Survey Guidelines
LBSC 878 Survey Guidelines
Spring 2005
The purpose of this page is to describe the survey that each student
in LBSC 878 will be asked to present. The survey is designed to
accomplish four goals:
- Dissertations include a survey of the related literature, and it
will be important to select an appropriate scope and organization for
that material. Developing a survey for this class will give you some
experience that should help when you start on your dissertation.
- Oral presentation is one of the most important skills that a new
Ph.D. will need. When looking for a job, an oral presentation is an
important part of the interview. When presenting your work at
conferences, oral presentations are the most prestigious and probably
the most influential format. And, of course, many Ph.D.'s are
employed in academia where classroom teaching is an important part of
the job. You cannot start too soon or practice too much if you seek
to become an effective speaker.
- You will need to prepare a reading list for your comprehensive
examination, and the development of a good reading list is one of the
first steps in doing a good survey.
- Our time together is limited, and there are practical limits on
how much reading each student can do outside of class. With your
survey, you will be able to introduce your classmates to ideas that
they might not have the chance to learn about in the depth that you
will understand it.
The mechanics of the survey are quite straightforward:
- First, choose a topic. This must be done by the time that class
meets in the second week. We will compare choices during class to
make sure that two people have not chosen the same topics. The class
Web page contains a list of suggested topics, but with the permission
of the instructors you may choose any subject that you find
interesting. Check with us before class if you want to pick a topic
that was not on our list, since we will need to be sure that it is
broad enough to introduce you to the issues that are discussed below.
- Once your choice of topics is approved, find one or more existing
surveys on that topic and read them. Your principal goal here is to
start the process of finding the relevant research literature (note:
research literature, not more surveys), so make a note of any
particularly promising references that you encounter in that first
survey.
- You should have time to read at least 10 research papers on your
chosen topic before presenting your survey (at 2 per week for 5
weeks). It is important to choose these papers well. Following
citation chains backward (using references you find) and forward
(using a citation index) are good techniques for finding good papers.
Browsing the latest conferences in the field and exploring the Web
pages of people who work in the area are good ways to find reports of
the latest work in the field. Keep your eye out for cognate fields
that might have materials you could productively bring in. Get
started on this as soon as possible, because some of the materials
that you need might have to come from Interlibrary Loan (which can
take about 2 weeks). Don't freeze your reading list at the beginning
- add things as you find them and then each time you sit down to read
something pick what looks to you to be the most productive reading on
the list at the time.
- Choose a point of view for your survey. A survey is essentially
a summary, and summarization only makes sense when done from some
perspective (otherwise, how would you decide what is NOT important?).
Another way to say this is that your survey needs a thesis - a claim
about the field that you will seek to substantiate. Since we will
read several surveys for class and you will start by reading at least
one survey of your chosen topic, you should have already seen several
examples of how to do this. In order to develop this skill, we insist
that you NOT choose the same point of view as the author of the first
survey that you read on your topic.
- Design an outline for your presentation. There are several
things to keep in mind:
- You will have at most 45 minutes for your talk, including
questions. It is better to plan a slightly shorter talk, since that
will leave time for questions along the way, extemporaneous thoughts
that you want to add, and the inevitable technical glitches. So plan
for about 30 minutes. A good rule of thumb is to prepare about one
slide for every 2 minutes, although speaking styles vary so much that
you will need to learn your own rate through experience. Plan on no
more than 18 slides unless you have some reason to believe that you
could get through more (or fewer) in 30 minutes.
- You will have the opportunity to assign one readings for your
audience to do beforehand, but you cannot assume that everyone
understood the reading(s) in the same way that you did. So your
survey should be self-contained, but perhaps covering issues that were
in your reading(s) more lightly. Note that you may not assign more
than one reading; it is your job to protect your classmates' time by
focusing their reading on the one paper that will help them to best
understand your survey. Note that this reading need NOT be a survey
paper. The most important thing is that it be a paper that causes
them to come to your talk with the right set of questions in mind.
- Every talk is a story, and it will help if you can think about it
that way. If your field has a long history, you might leverage that
my presenting the talk as a history lesson that leads up to the
present state of the art. If your field integrates techniques from
many disciplines, you might choose one discipline and then show how
the others have affected it. If there are competing perspectives from
which people work on the field, you might explain each and then argue
for one of them.
- Use examples to illustrate complex points, either before or after
presenting the general issue.
- Present an outline at the beginning, and then refer to it at
major points along the way so that your audience can see where you are
with respect to the big picture.
- About two weeks before your talk, make an appointment with one of
the instructors. This appointment is not optional -- you must make
one even if you believe that you are well prepared. At this point you
should have done most of your readings and have completed an initial
draft or at least an outline of your talk. Bring you're your draft
talk and your entire reading list (indicating which ones you have
actually read) to the meeting. Our goal will be to go over everything
that you have done and then help you to make the best use of the
remaining two weeks.
- When you have a small audience (as you will), it is sometimes
useful to hand out materials before you speak. If you want the
audience to pay attention to you rather than taking notes, handing out
reduced copies of your slides can be helpful (Powerpoint will put 2-9
on a page, depending on version and the size you want). If you are
concerned with the audience looking ahead, you might only distribute a
few selected slides. This can be a particularly good idea if some of
your slides have too much detail to be understood at a glance
(althought you should avoid that anyhow when designing your slides).
You may create any form of handout you like, or no handout at all, at
your option.
- When you actually make your slides, keep them simple! This is
hard to do, but very important. Use illustrations whenever they can
be as effective as text, because people can process moderately complex
images much more easily than they can process moderately complex text.
Don't try to get more than one major idea on each slide. When using
text, stick with Powerpoint's defaults. Don't make lines smaller.
Don't make bulleted items more than one line. Remember that you will
be talking - that's why we call it a talk! The purpose of the slides
is to reinforce what you are saying, not to say it for you.
- Make a script if you think you need one. There are some
disadvantages to using a script:
- Using a script makes it more difficult to make eye contact with
your audience. Eye contact is important for two reasons. First, it
makes your story more compelling and thus holds your audience's
interest better. Second, and perhaps more important, you can get a
sense for how the audience is reacting. That way you can go back to
basics and emphasize your main point if everyone looks lost, or pick
up the pace a bit if everyone looks bored.
- Using a script makes your message seem less personal. After all,
couldn't anyone have given your talk once the script was available.
Your audience wants to hear you, not just your message.
- Preparing a complete script takes a lot of time. Everything
about preparing a talk takes time, so you will have to make
compromises somewhere. There are lots of better places to invest the
time you have -- particularly in the organization of your talk.
Of course, many people do prefer to use a script despite these
disadvantages. If you do choose to make one, there are a few things
to bear in mind:
- An outline of your remarks can be better than a word- by-word
script because it will result in a more natural delivery and can be
prepared more quickly.
- Remember that you will have to split your attention between the
script and the audience. Use a much larger font than normal and
double space the script. This helps to avoid losing your place.
- Align your script with your slides so that every time you flip a
slide you flip a page on the script. This will help you avoid getting
lost in the script. Powerpoint has a "notes" feature for doing this.
- Practice! For the first few talks that you give, you should
conduct a minimum of two, and preferably three, practice sessions.
The first practice should be alone, and your goal should be to measure
the time required to deliver your talk at a natural cadence and to
find any rough edges in what you plan to say. This is known as
"talking to the wall," and it is remarkably useful. The second
practice talk should be given to at least one person, even if they do
not have the background to understand your subject. They can help you
with your delivery, and you will be able to spot some more rough
edges. Give them a copy of your talk and ask them to mark the time
that you start each slide. That will help you to understand where you
are investing time if (as is often the case) you are still running
long at this point. Ask them to be candid in their comments.
Finally, try to get a group together that would understand your talk
and give the talk to them. Don't use your LBSC 878 classmates,
though, or you will spoil the impact of your real talk to them. The
goal here is to discover problems with the organization of your talk
or the details of your presentation. This step is optional because
time and the availability of an audience sometimes makes this
impractical.
- Present the talk in class. It is helpful to be as relaxed as
possible at this point. The best way to be relaxed is to be confident
that you are well prepared, and that is what all of the work up until
now has been for. You should come at least 15 minutes early if you
will be the first speaker in order to ensure that the technology is
working. When speaking, shift your gaze frequently from one member of
the audience to another. In a large group, it helps to pick one
person on your left, one in the center, and one on your right to talk
to. But in a small group you should talk to each person at least some
of the time. Whatever you do, don't just talk to the professors!
Inform your audience before you start if you want questions as you go
along, or whether you want all questions held to the end. A good
compromise here is to allow only clarifying questions as you go along
-- otherwise you can get into long debates that would make it
impossible for you to get to your conclusion.
- Take questions at the end. Surprisingly, it is common to see a
student give a great talk that is followed by a weak question period.
That often results from too much time spent polishing the speech and
not enough time spend really thinking about your subject. It is not
important that you know everything -- that would be impossible! If
you get a question that you don't know the answer to, say so. If a
questioner points up a limitation of your analysis, acknowledge it.
This is an opportunity for your audience to engage an expert in a
discussion, and you are that expert. Be ready to fill that role.
There are two good ways to prepare for questions. One is to seed some
questions in the audience's mind by leaving some interesting but not
essential details out of your talk, perhaps mentioning that "we can
discuss this further in the questions if you like." Of course, don't
do that so much that it disrupts the flow of the talk. The other
thing you can do to prepare is to have your practice audience ask you
some questions at the end (if you have a practice audience). That
will give you some practice thinking on your feet.
- Meet with both instructors as soon after your talk as possible.
Right after class would be best. Depending on schedules, we can do
this either all at once or separately. If time allows, we may make
some comments in class that we think would be of general interest, but
we'll usually also have some tips or observations to discuss with you
privately. This sort of feedback is hard to get from a real audience,
so it can be one of the most valuable parts of doing this in a
academic setting.