I'm trying to understand how to graph percent changes in Stata. (As a visual reference, I'm trying to do something very similar found here on page 436:
http://atop.rice.edu/download/public...eedsAJPS03.pdf)
I have a dependent variable that is a rate (percentage bound by 0-100) and the unit of analysis is directed dyad year (country a and country b year/ country b country a year). I want to make a graph showing the percentage change of some of the independent variables (some of which are binary), with all other variables being held at their means, but I'm not even sure where to start.
That being said, I guess my first question is less technical. Is this kind of graph even possible ? Or is it more geared if your dependent variable is binary? Is there a better way to graph percent changes of independent variables given Stata's graphing options? I tried to look on statalist but the most I could get was a line graph that was asked here:
http://www.statalist.org/forums/foru...ver-line-graph
In order to create more options, also calculated the percent change of the actual dependent variable using
where "smoothtotrade" is the trade between country a and country b in a given year.
Essentially, what I'm trying to show is the change in the rate given, (for example), the amount of trade between two countries, with all other variables held at the means. Is the percent change over the line graph a better option? I'm trying to visualize the data besides the standard regression table. My understanding is that
command is more for partial effects/logit.
Lastly, how do you hold other variables at their means in Stata? Is it just
?
Regards
http://atop.rice.edu/download/public...eedsAJPS03.pdf)
I have a dependent variable that is a rate (percentage bound by 0-100) and the unit of analysis is directed dyad year (country a and country b year/ country b country a year). I want to make a graph showing the percentage change of some of the independent variables (some of which are binary), with all other variables being held at their means, but I'm not even sure where to start.
That being said, I guess my first question is less technical. Is this kind of graph even possible ? Or is it more geared if your dependent variable is binary? Is there a better way to graph percent changes of independent variables given Stata's graphing options? I tried to look on statalist but the most I could get was a line graph that was asked here:
http://www.statalist.org/forums/foru...ver-line-graph
In order to create more options, also calculated the percent change of the actual dependent variable using
Code:
gen percentchange = (lag_rate/lag_rate[_n-1] - 1)*100
Code:
input float(percentchange smoothtotrade)
. 398258.8
-26.2685 375342.8
22.22222 366923.25
-5.333336 387282.1
-83.00291 436064.7
Essentially, what I'm trying to show is the change in the rate given, (for example), the amount of trade between two countries, with all other variables held at the means. Is the percent change over the line graph a better option? I'm trying to visualize the data besides the standard regression table. My understanding is that
Code:
margins
Lastly, how do you hold other variables at their means in Stata? Is it just
Code:
rest (mean)
Regards
