When I try to use fullmatching like so:
> a <- matchit(glmForm, nData, method = "full", min.controls = 1, max.controls = 5)
And then try to look at the matches like so:
> a$match.matrix
> I get a response of NULL, indicating that the object doesn't exist. However when I use method = "nearest" or method = "optimal" it works just fine. How can I see the matches that are produced using full-matching? As an arbitrary example, I need to be able to see that patient 1 in the treatment group is matched to patients 4,5,34,31,and 16 in the control group, and patient 2 in the control group is matched to patients 7,23,523,89,and 24 in the treatment group.
Thanks,
Skye Buckner-Petty
Statistical Data Analyst
GMS
Department of Medicine
Office - (314) 362 4185
FAX - (314) 454 5113
Cell - (314) 732 3030
HIPAA: The materials in this e-mail message are private and confidential and may also contain protected healthcare information. If you are not the intended recipient, be advised that any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this e-mail in error, please immediately notify the sender via telephone or return e-mail.
Full matching is a subclassification procedure and so you should look at a$subclass
Kosuke
Department of Politics
Princeton University
http://imai.princeton.edu
On Nov 27, 2012, at 1:02 PM, Buckner-Petty, Skye wrote:
> Hi Kosuke,
>
> When I try to use fullmatching like so:
>
>> a <- matchit(glmForm, nData, method = "full", min.controls = 1, max.controls = 5)
>
> And then try to look at the matches like so:
>
>> a$match.matrix
>
> I get a response of NULL, indicating that the object doesn't exist. However when I use method = "nearest" or method = "optimal" it works just fine. How can I see the matches that are produced using full-matching?
>
> Thanks,
> Skye
> -----Original Message-----
> From: Kosuke Imai [mailto:kimai@Princeton.EDU]
> Sent: Wednesday, November 07, 2012 7:13 PM
> To: Buckner-Petty, Skye
> Cc: <matchit(a)lists.gking.harvard.edu>
> Subject: Re: R MatchIt package
>
> Yes, the variable has to be 0-1 I think.
>
> Kosuke
>
> Department of Politics
> Princeton University
> http://imai.princeton.edu
>
> On Nov 7, 2012, at 1:53 PM, "Buckner-Petty, Skye" <SBUCKNER(a)dom.wustl.edu> wrote:
>
>> Hi Kosuke,
>>
>> I am doing some analysis that requires matching and recently come across your package. However, when trying to run the following code, I get an error:
>>
>>> bbb= matchit(formula = compLbl ~ Age + Race_Code + Complexity_Subclass_Code, data = nData, method = "nearest")
>> Error in weights.matrix(match.matrix, treat, discarded) :
>> No units were matched
>> In addition: Warning messages:
>> 1: In max(pscore[treat == 0]) :
>> no non-missing arguments to max; returning -Inf
>> 2: In max(pscore[treat == 1]) :
>> no non-missing arguments to max; returning -Inf
>> 3: In min(pscore[treat == 0]) :
>> no non-missing arguments to min; returning Inf
>> 4: In min(pscore[treat == 1]) :
>> no non-missing arguments to min; returning Inf
>>
>> The variable "compLbl" is dichotomous (group A and group B), but it is not binary coded (1s and 0s). Does the treatment variable need to be binary coded in order for it to work?
>>
>> Thanks,
>>
>> Skye Buckner-Petty
>> Statistical Data Analyst
>> GMS
>> Department of Medicine
>> Office - (314) 362 4185
>> FAX - (314) 454 5113
>> Cell - (314) 732 3030
>>
>>
>> HIPAA: The materials in this e-mail message are private and confidential and may also contain protected healthcare information. If you are not the intended recipient, be advised that any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this e-mail in error, please immediately notify the sender via telephone or return e-mail.
>>
>>
>>
>
Yes, the variable has to be 0-1 I think.
Kosuke
Department of Politics
Princeton University
http://imai.princeton.edu
On Nov 7, 2012, at 1:53 PM, "Buckner-Petty, Skye" <SBUCKNER(a)dom.wustl.edu> wrote:
> Hi Kosuke,
>
> I am doing some analysis that requires matching and recently come across your package. However, when trying to run the following code, I get an error:
>
> > bbb= matchit(formula = compLbl ~ Age + Race_Code + Complexity_Subclass_Code, data = nData, method = "nearest")
> Error in weights.matrix(match.matrix, treat, discarded) :
> No units were matched
> In addition: Warning messages:
> 1: In max(pscore[treat == 0]) :
> no non-missing arguments to max; returning -Inf
> 2: In max(pscore[treat == 1]) :
> no non-missing arguments to max; returning -Inf
> 3: In min(pscore[treat == 0]) :
> no non-missing arguments to min; returning Inf
> 4: In min(pscore[treat == 1]) :
> no non-missing arguments to min; returning Inf
>
> The variable “compLbl” is dichotomous (group A and group B), but it is not binary coded (1s and 0s). Does the treatment variable need to be binary coded in order for it to work?
>
> Thanks,
>
> Skye Buckner-Petty
> Statistical Data Analyst
> GMS
> Department of Medicine
> Office – (314) 362 4185
> FAX – (314) 454 5113
> Cell – (314) 732 3030
>
>
> HIPAA: The materials in this e-mail message are private and confidential and may also contain protected healthcare information. If you are not the intended recipient, be advised that any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this e-mail in error, please immediately notify the sender via telephone or return e-mail.
>
>
>