I have corrected the typos in the lecture notes mentioned in class.
See: http://jsekhon.fas.harvard.edu/
Cheers,
JS.
======================================
Jasjeet S. Sekhon
Assistant Professor
Harvard University
Center for Basic Research in the
Social Sciences
jsekhon(a)fas.harvard.edu
http://jsekhon.fas.harvard.edu/
Office: 617.496.2426 Fax: 617.496.5149
======================================
greetings. some questions from some of us on the problem set....
we have some questions about the Pepsi problem, part b. Are people assuming that the question is asking what the probability that any one person, anybody, or any of the ten people selected has picked the same number?
we were looking at the conditional probability that you have the right number, given that you were one of the ten. using bayes, this seemed to be 1 in 10,000, though if you find the probability that ANY of the finalists have the number it is 1 in 1000.
the whole thing seems problematic because you have to account for the people who have 3, 4, 5 and 6 matching numbers in the final ten (though the probability of any of these people existing is pretty trivial), and thus the different mix of possibilities in the group of 10.
any help is appreciated...
best,
jason
An updated version of the lecture notes is available online
(http://jsekhon.fas.harvard.edu/). Today's lecture will begin on page
50 ("Independence and Experimental and Nonexperimantal Research"). I
plan to go on to cover the Bernoulli, binomial, and normal
distributions.
Cheers,
JS.
======================================
Jasjeet S. Sekhon
Assistant Professor
Harvard University
Center for Basic Research in the
Social Sciences
jsekhon(a)fas.harvard.edu
http://jsekhon.fas.harvard.edu/
Office: 617.496.2426 Fax: 617.496.5149
======================================
The online lecture notes have been updated. Please download the new
versions. Today's lecture will begin on page 50 "Independence and
Experimental and Nonexperimantal Research". I hope to cover the
Bernoulli, binomial and normal distribution and get to page 80.
Cheers,
JS.
======================================
Jasjeet S. Sekhon
Assistant Professor
Harvard University
Center for Basic Research in the
Social Sciences
jsekhon(a)fas.harvard.edu
http://jsekhon.fas.harvard.edu/
Office: 617.496.2426 Fax: 617.496.5149
======================================
Hi all -
I have a question about the discussion we had tonight regarding the
confidence interval versus the sample confidence interval. I understand
that when you have to estimate both the mean and the standard deviation to
calculate the confidence interval, you have to use n-1 rather than n.
The formula that we are using to calculate the confidence interval only
requires us to estimate one parameter, p. Using this method for
calculating confidence intervals, when would we need to use n-1?
Cheers,
Mike
Hello Everyone,
Because the course is relatively large (35 enrolled students at last
count), I am going to have to turn away some auditors. I apologize
for the inconvenience but the first priority must be the students who
are taking the course for credit. If you can present a *VERY* good
reason for auditing the course, I may allow you to stay.
Note 1: I have been very pleased with the discussion on the mailing list.
You're asking great questions and posting wonderful responses.
Note 2: There will be no class October the 6th because of Yom Kippur.
Note 3: I would like meet with all of the enrolled students
(one-on-one) over the next couple of weeks. I will set aside extended
office hours Wed the 1st and 8th for this purpose. I will be
available from 2 to 7pm on both days. Jacob will send around sign-up
sheets on Monday.
Cheers,
JS.
hey everyone,
i swear i understood this five minutes ago when it was explained to me
but now it has escaped again....it always seems to come back to this
problem of overlapping confidence intervals - for part b) of the first
question, how do we show that 52% is lower than 55% with 95% confidence
when the 95% confidence interval for 55% seems to include 52%? or have i
done the confidence interval wrong?
(i'm sorry if this question has already been asked - i haven't been
getting the list's emails until tonight).
thanks,
zoe
My apologies: ISwR and not IswR. If ts is already on your computer, then
you are taken care of.
On Tue, 23 Sep 2003, Amit Rajendra Modi wrote:
>
> Hi Jacob,
>
> A couple questions about installing R. I followed the instructions on the
> website, and the first and third commands work fine. However,
>
> install.packages("IswR", .libPaths()[1])
>
> gives the message : No package "IswR" on CRAN. in: download.packages(pkgs,
> destdir = tmpd, available = available,
>
> and the command
>
> install.packages("ts", .libPaths()[1])
>
> gives the message:
>
> Warning message:
> package ts is in use and will not be installed
>
>
> How do I get around these? Is it necessary to get around them?
>
> Thanks,
> Amit
>
> Quoting Jacob Kline <jkline(a)fas.harvard.edu>:
>
> >
> > To receive emails related to the course, you need to subscribe to the email
> > list gov1000-list@fas. Do so using the following link:
> >
> > http://www.fas.harvard.edu/mailman/listinfo/gov1000-list
> >
> > When you have a question about the course, mail your question directly to
> > this list and not to the course instructors. (We will receive them from the
> > list.) If you feel that you are able to answer someone's question, feel free
> > to do so. The course instructors will be monitoring to make sure that correct
> > information is being propogated.
> >
> > Hope you are well,
> > Jacob
> >
>
>
>
>
Based on today's discussion of distributions, do we think Plan B is better than Plan C in problem 2? Based on the initial distribution, these distributions appear to be the random distributions that are most similar to the original distribution, even though the numbers in the two groups are different. i think the distribution is more important than the total number of cases...other thoughts?