Hi Everyone,
I just put a slightly revised (2 readings added) syllabus up on the
course website.
Best,
Kevin
------------------------------------------------------
Kevin Quinn
Assistant Professor
Department of Government and
Center for Basic Research in the Social Sciences
34 Kirkland Street
Harvard University
Cambridge, MA 02138
Hi Everyone,
I just ran across an application of regression analysis that, while
not really social scientific, illustrates a number of concepts we'll
be talking about in the next lecture. The cite is:
Tatem, Guerra, Atkinskon, and Hay. 2004. "Momentous Sprint at the 2156
Olympics." Nature. 431 (30 September) 431: 525.
The article and data are up on the course website under the handouts
section.
Please read this and spend a bit of time thinking about why you do or
do not believe their results. We'll talk about this at the next
lecture.
Best,
Kevin
------------------------------------------------------
Kevin Quinn
Assistant Professor
Department of Government and
Center for Basic Research in the Social Sciences
34 Kirkland Street
Harvard University
Cambridge, MA 02138
> I'm having trouble even knowing where to begin for the coding on these
> problems and because I'm not grasping the stats yet, the section 2
> houndout is unfortunatly not helping me much. I did get the Fredman
> stats book, so that's definitely helping the more I read. But the main
> issue is coding and knowing what the output is supposed to be/look like.
> Any tips are great!
Gov1000 friends: Below is part of a message from the VNC script author.
When you download the script subsequently, you'll need to enter the
hostname of a specific iceX machine. Let me know if you have any
questions.
Ryan
---------- Forwarded message ----------
Date: Wed, 29 Sep 2004 17:12:48 -0400
From: Erik Ray via RT <Director_Support(a)help.hmdc.harvard.edu>
Therefore, I have removed "nice..." from the hostname query box and left
it blank.
> Hi Ryan,
>
> the homework is going pretty well, and I have one minor question: When we
> are working with Z ~ N(0,1), and are asked things like Pr(|Z|>c), I just
> calculate the Pr(Z>c) and multipy by 2, relying on the symmetry of
> the normal distribution along either side of the y-axis. I am assuming
> this gives me a correct answer, but would there be a way to actually work
> with |Z| in R?
The symmetry of the normal curve allows exactly the sort of assumption
you've made. Note that for a different member of the Normal family, like
N(4,1), the symmetry would not be around the y-axis, but x=4.
> R does not seem to want to let me type pnorm(|1|, lower.tail=FALSE). Is
> there something else I should try?
The R function "abs(x)" returns the absolute value of x. Feel free to use
it if you like.
Cheers,
Ryan
Yes, it's me again. Just giving a heads up to people that me and others
have decided to meet at Friday at 2 in the Science Center computer lab
(basement) to work on the 1000 problem set. I'm not sure what other
study groups are already formed, but do feel free to come.
In my econ study group, we found that it was useful if people worked on what
they could in the problem set before meeting. I guess in terms of this
class this translates to not just the problems themselves but also the
computer programming, since (esp for the first problem set) figuring out
Latex and R is atleast half of the work.
Let me know if you have any questions.
Best,
Becky
Good news for those of you who still need to buy the Cleveland book--
see below from the COOP...
Best,
Kevin
---------- Forwarded message ----------
Date: Wed, 29 Sep 2004 15:05:20 -0400
From: tm300(a)bncollege.com
To: Kevin Quinn <kquinn(a)fas.harvard.edu>
Subject: Re: Visualizing Data
The 35 copies of Visualizing Data just arrived. We are pulling copies for
customers who reserved them and shelving the rest.
Sarah
Hi-
Sorry to clog email boxes some more, but I am wondering if anyone is
having success importing R graphs into Latex. I successfully saved my
graphs from R as pdf's. The pdf's are saved in teh same folder as my
Latex document. Here is my code I used (copied from
directly from the hwtemplate.tex document) to import my graph into Latex:
\begin{figure}[h]
\centering
\scalebox{.7}{\includegraphics{graph1.pdf}}
\caption{}
\label{f1}
\end{figure}
I use the same code copied two more times for three different graphs
(graph1.pdf, graph2.pdf, graph3.pdf)
This is the error message I get when I try to complie Latex in the
Terminal screen:
! LaTeX Error: Cannot determine size of graphic in graph1.pdf
See the LaTeX manual or LaTeX Companion for explanation.
Type H <return> for immediate help.
1.65 ...scalebox{.7}{\includegraphics{graph1.pdf}}
?
<graph1.pdf>
Then the error repeats 2 more times, once for each graph. When I open
the .dvi document, I only see the first graph rotated on its side, some of the
subsequent text is missing, and then it looks normal starting on the
following page.
Any ideas? Ryan, did you say you had office hours following section
tomorrow?
Thanks,
Becky
Hi all-
We just wanted to make sure that everyone is clear on expectations for the
homeworks.
HOMEWORK GUIDELINES
-homeworks must be "sufficiently started" the first time they are turned
in to receive credit. This mean that the student will have startedand
made significant progress with each
problem. The instructors have the discretion to judge what constitutes
"sufficiently started" in individual cases.
-please include your R code at the end of your problem set. Also, if you
work in groups, you must write your own code (i.e., type in commands
yourself) rather than cutting and pasting code written by peers
-homeworks must be type-written (this makes re-writes easier), preferrably
in LaTeX
-if you worked in a group, list your group members at the top
-please estimate how long the problem set took you and put this at the top
-please star any problems that you found to be particularly
informative/interesting
-homeworks must be turned in at the beginning of class on Mondays and must
be printed out rather than emailed
Please let us know if any of this is unclear!
Kevin, Ryan and Alison