Maya,
Thank you. Does that mean if my coefficient estimate for the contact
variable is 0.7262, I interpret it as: the probability of turnout for
someone who is contacted is about (0.7262*0.4) larger than the
probability of turnout for someone who is not contacted ?
Thanks!
Iza
On Apr 8, 2010, at 5:08 PM, Maya Sen wrote:
Hi Iza,
Gary discussed in class some quick heuristics that you can use to
interpret the "effects" of both Logit and Probit coefficients.
Specifically, the max value for a logit "effect" is going to be
beta/4, while for a probit, the max value will be beta * 0.4.
If what you're looking for are fitted values, then you can use
z.out$fitted.values where z.out is your zelig object. (There's
something similar in glm.) Otherwise, I think you are looking to use
zelig's setx and sim commands together to simulate quantities of
interest from the posterior distribution. There are some good examples
in the zelig manual.
hope that helps,
Maya
On Thu, Apr 8, 2010 at 5:02 PM, Iza Ding <yding at fas.harvard.edu>
wrote:
Hey Class,
I have some trouble interpreting the Probit coefficient estimates:
what does one unit's increase in the value of a given variable lead
to? Also, what is the R function for predicted probabilities from the
Probit outcome?
Many thanks.
Iza Ding
_______________________________________________
gov2001-l mailing list
gov2001-l at
lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
_______________________________________________
gov2001-l mailing list
gov2001-l at
lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l