I ran three logit models in R with the Zelig package and I'm trying to
compute the
predicted probabilities for a number of different values on the independent
variable.
My dep variable was accepted or decline and my indep variable is bid
amount, and varies.
So for a bid amount of 3, what's the expected probability of winning.
For a bid amount of 5, what's the expected probability of winning.
and so forth.
I ran the following code, but I'm getting some weird results.
d1 = data.frame(mwin=c(mwin), mbid=c(mbid))
str(d1)
m1 = zelig(mwin ~ mbid, data=d1, model="logit")
summary(m1)
plot(mwin ~ mbid)
> str(d1)'data.frame': 66514 obs. of 2 variables:
$ mwin: int 2 2 2 2 2 2 2 2 2 2 ...
$ mbid: int 700 300 700 300 500 300 300 700 300 300 ...
library(Zelig)
x.out <- setx(m1, mbid=300)
s.out <- sim(m1, x = x.out)
summary(s.out)
plot(s.out)
When I run with mbid as 300, I get 49%.
At 500, it's 49% and at 700 it's 50%.
At 1500, it's 51%
These results are just really weird.
I was expecting an exponential curve when I plotted
mbid by probability of winning, but that doesn't seem
to be the case
--
*Abraham Mathew
Statistical Analyst
www.amathew.com
720-648-0108
@abmathewks*
[[alternative HTML version deleted]]
