Hi Ian
So sorry to bother you again. Actually this stuff is so important that I
would like to make sure that I understand it properly. I dont know why I
have problem understanding it; perhaps due to non familiarity with various
notations or matrix algebra, whatever, but I really want to understand it
properly and that is why I am giving you some problem. I really appreciate
your help on this.
Okay! I have gone through the notes and the article and every thing but I am
still not clear. What do you mean when you say "you can use the draws of the
coefficient to simulate uncertainty about these fitted values" or when Gary
says in his article on page three and point 2 "Draw one value of the
vector .... ...from the multivariate normal distribution in Equation 4.
Denote the ......."
and points 3 nad 4 on the same page
3. Taking the simulated effect .......
4. Simulate the outcome variable Y hat........
Can you give me an example of two or three random draws using the Logit
model from Gary,s article the one he simulated from NES study by
Rosenstone and Hanson. What I really did not understand was *how he repeated
for each case the expected value algorithm M = 1000 times to approximate a
99 percent confidence intervals around the probability of voting.*
I would reaaly appreciate your help. Thanks again
Bilal
On 3/28/06, Ian Brett Yohai <yohai(a)fas.harvard.edu> wrote:
Hi Bilal,
If you look at the section7 handout (in the Sections folder on the course
website), there are a few examples that does what I think you would like.
See subsection 4 called "R code" and also subsection 5 which shows how
the Zelig syntax works for this type of thing.
In your particular example, I don't see a 'race' term in your logit
equation. You can estimate a logit with income, race, and education on
the right hand side. Then when you simulate a first difference, you can
hold income and education at their means, while changing race presumably
from a zero to 1 if you have race coded as a dummy variable. In Zelig
this is achieved by using two setx commands - again, see subsection 5 of
the section 7 handout.
On your last point about changing from 45% to 50% turnout, I'm not sure I
follow entirely. But you can play around by changing the levels of
education in your first difference setup, while holding the other
variables at means (or at some other values that you deem substantively
important) and computing the effect on turnout.
You need to take draws from a multivariate normal distribution to make
this work, so I definitely would suggest moving to R.
Best,
Ian
On Tue, 28 Mar 2006, Bilal Khan wrote:
Hi All
Can somebody help me to understand the two types of simulation that Gary
gave lecture on. I am still bit confused. I use SPSS for my logit works
but
I strongly believe that we have to move beyond
calculating simple betas
and
odds and give quantities of interest along with
uncertainity.
Suppose Beta = .0250I for education and Beta = .06531 for income in a
logistic regression equation: Logit (turnout) = .02501 education +
.06531
income. I would like to know through an example
how would you simulate
the
impact of race on turnout
1. while holding constant income and education at their means.
2. with income bracket of 30,000 to 45,000 dollars and less than high
school
of education.
Can somebody give example by drawing three to four samples?
Also many times when you have predicted probabilities of voting in an
election for a data set using logistic regression model for each case in
the
sample of a state or an area and after
considering probability of less
than
.50 not voting and more than .50 voting, how can
you show the impact of
changing a value of the parameter e.g. education with less than high
school
to all the sample having atleast high school
education, on the predicted
turnout of say 45 percent for the sample.
That is I would like to say that changing a certain parameter (kind of
first
difference) the total turnout would improve from
45 percent to 50
percent or
whatever.
I know I can do that in SPSS but it wont give me uncertanity or
confidence
intervals: which most of the analysts dont give
for such type of "what
if
analysis" I am going through the work of
Wolfinger and Rosentone "Who
votes"; excelent work but no confidence interval levels or uncertanity
in
explaining their quantities of interest
claculating through probit.
How can you use Zelig for producing such quantities of interest?
Bilal
_______________________________________________
gov2001-l mailing list
gov2001-l(a)lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l