Hi there,
I'm wondering if Matchit package allows to match on the true propensity score - meaning that instead of estimating PS (what matchit function does) I would be able to provide a vector of true propensity scores which would be then used in the process of matching. The second thing that I'm interested in is, whether it is possible to do Mahalanobis distance matching and not only propensity score matching. I did check your documentation (perhaps not carefully enough) but couldn't find my answers there.
I'm looking forward to hearing from you soon!
Ana
Yes, you can do nearest neighbor matching with exact restrictions. The option is called "exact": see http://gking.harvard.edu/node/4355/rbuild_documentation/Additional_Argument… For optimal matching, you would have to tweak the inputs to outmatch package by Ben Hansen. I'm not sure how exactly you would do that, but if you figured it out, then you can pass those arguments through matchit() function.
Kosuke
Department of Politics
Princeton University
http://imai.princeton.edu
On Sep 14, 2012, at 5:22 PM, Joshua Stewart wrote:
> Kosuke Imai,
>
> I am writing in regard to a question I have with your MatchIt R program. Currently I am trying to analyze a data set with continuous and categorical data. We have achievement score of students, race, school ID, free and reduced lunch status and treatment or control status. My question relates to whether or not it is possible to perform an exact match on categorical data such as race, while at the same time run and optimal or nearest neighbor match on achievement scores. Thank you very much for your time.
>
> Regards,
>
> Joshua Stewart
> Researcher
> Mid-continent Research for Education and Learning (McREL)
> 4601 DTC Blvd, Office 480
> Denver, CO 80120
> 303.632.5508
> www.mcrel.org
>
I would like to replicate the scaling factor of the dots in the matchit jitter
plot for data points that are weighted for a modified graph.
I am estimating the ATT, so all treated subjects are equal to cex=1.
In the documentation, it says "the size of each point is proportional to the
weight given to that unit".
Is the cex function the appropriate way to scale the size of each
point?
For example, if a control subject is matched with a weight of 2.2, would
that mean the appropriate point size is cex=2.2? Similarly, if the control
subject weight is 0.25, then the cex=0.25?
When I do this, it does not seem to be the same scaling factor as the MatchIt
jitter plot (the larger points look too big and the smaller points look too
small). I also could not find the source code that generates the standardized
jitter plot (is that available?).
Any help is very appreciated.