Open christophergandrud opened 7 years ago
I guess a temporary workaround is that zelig_qi_to_df
returns the simulations for all of the by
models, e.g.:
library(Zelig)
library(dplyr)
zelig(Petal.Width ~ Petal.Length, by = "Species", data = iris,
model = "ls") %>%
setx(z.5, Petal.Length = 2:4) %>%
sim() %>%
zelig_qi_to_df() %>%
head()
## setx_value by Petal.Length expected_value predicted_value
## 1 x setosa 2 0.3711907 0.4449542
## 2 x setosa 2 0.3863161 0.2804666
## 3 x setosa 2 0.3756340 0.3544871
## 4 x setosa 2 0.4040592 0.3268610
## 5 x setosa 2 0.3811405 0.2781935
## 6 x setosa 2 0.3657079 0.3243758
Do you have an idea of what ideal plot behavior would be. Presently, we build up this panel of very many different plots, showing different ways of interpreting whatever qi's exist. Under that present logic, would you create all of these panels, and cycle through them? Create some higher level summary? Have a default panel to show, but allow the user to somehow specify any of the other by
conditions?
(My sense is perhaps the latter makes sense, and then if a user wants to see all the panels, they cycle through the plot
calls)
I initially thought of just cycling through. But maybe an overview would be good. This could be easy with ci plots. Just put them in a grid.
I am trying to obtain att.ev from sim output but i keep getting an error, please see code below:
z.m_b1 <- zelig(log(wy_yield_plot+1) ~ hhead_age + hh_education_mean + xx55 + w_c + perc_own_title + bean_area + dem, data=out_data_m_b1_control,model="ls")
x.out1 <- setx(z.m_b1,data=match.data(m_b1="treat"),cond=TRUE)
s.out1 <- sim(z.m_b1, x=x.out1)
Here i tried different ways to obtain att.ev values from sim output: s.out1$qi$att.ev Error in s.out1$qi$att.ev : object of type 'closure' is not subsettable
then i tried
s.out1\$qi\$att.ev
Error: unexpected input in "s.out1\"
Please advice, Thanks so much for your support.
Thanks for your issue @jefunes. This seems like a separate topic so I created a new issue over at: https://github.com/IQSS/Zelig/issues/243
Please follow progress there.
In the following code, the model is estimated
by
whether the canton is majority or minority catholic. However the plots are only created for one model estimation.