I tried the past two days getting psmatching to operate properly. I also couldn't find the right solution in the discussion topics as I run through it. Maybe you can help me out?
I installed PSMATCHING3.03 on the fixed IBM SPSS 21.0.0.2, R version 2.14.2, and R essentials for spss 21 on windows xp system 32 bit.
If I try Optimal Matching via the following syntax:
PSMATCHING3
/VARS
ID = ID
TREAT = Controlled_vs_Uncontrolled
COVS = weight length followupDuration
/MATCHIT
MATCH=OPTIMAL
EST =LOGIT
DISCARD = NONE
MORDER = LARGEST
RATIO = 1
CALIPER = 0
/PLOT HISTPLOT JITTERPLOT HISTBAL DOTPLOT INDBAL RESOLUTION = 150
/OUTPUT ALL MATCHED.
I get only this output when running the syntax above:
Error
fullmatch unused argument(s) (data = data)
Error in xy.coords(x, y, xlabel, ylabel, log) :
'x' is a list, but does not have components 'x' and 'y'
Nearest neighbor matching seems to work fine, however, Fullmatching doesn't work either (and Gam logit doesn't work also). The program seems to run fine in the beginning, but it ends abruptly as stated above in the output. Why does only nearest neighbor matching seem to work, and how can I make use of optimal matching and GAM-logit? Does it has to do something with new updates of R?
Thanks in advance for your answer.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I tried the past two days getting psmatching to operate properly. I also
couldn't find the right solution in the discussion topics as I run through
it. Maybe you can help me out?
I installed PSMATCHING3.03 on the fixed IBM SPSS 21.0.0.2, R version 2.14.2,
and R essentials for spss 21 on windows xp system 32 bit.
If I try Optimal Matching via the following syntax:
PSMATCHING3
/VARS
ID = ID
TREAT = Controlled_vs_Uncontrolled
COVS = weight length followupDuration
/MATCHIT
MATCH=OPTIMAL
EST =LOGIT
DISCARD = NONE
MORDER = LARGEST
RATIO = 1
CALIPER = 0
/PLOT HISTPLOT JITTERPLOT HISTBAL DOTPLOT INDBAL RESOLUTION = 150
/OUTPUT ALL MATCHED.
I get only this output when running the syntax above:
Error
fullmatch unused argument(s) (data = data)
Error in xy.coords(x, y, xlabel, ylabel, log) :
'x' is a list, but does not have components 'x' and 'y'
Nearest neighbor matching seems to work fine, however, Fullmatching doesn't
work either (and Gam logit doesn't work also). The program seems to run fine
in the beginning, but it ends abruptly as stated above in the output. Why
does only nearest neighbor matching seem to work, and how can I make use of
optimal matching and GAM-logit? Does it has to do something with new updates
of R?
Hello,
After consulting with my grad student who helped programming, we
believe that the error that optimal / full matching are not working in
your case, might be that the caliper is set to 0. The error message is
indeed not very clear, but if you can change that value and check if
it still not works, let us know. If the caliper is set to 0, only
identical matches may be allowed, which will often result in no
matches at all.
As to whether GAM is not working, I am not sure and we will look into
this. It could also be related to the caliper and less to GAM.
I tried the past two days getting psmatching to operate properly. I also
couldn't find the right solution in the discussion topics as I run through
it. Maybe you can help me out?
I installed PSMATCHING3.03 on the fixed IBM SPSS 21.0.0.2, R version 2.14.2,
and R essentials for spss 21 on windows xp system 32 bit.
If I try Optimal Matching via the following syntax:
PSMATCHING3
/VARS
ID = ID
TREAT = Controlled_vs_Uncontrolled
COVS = weight length followupDuration
/MATCHIT
MATCH=OPTIMAL
EST =LOGIT
DISCARD = NONE
MORDER = LARGEST
RATIO = 1
CALIPER = 0
/PLOT HISTPLOT JITTERPLOT HISTBAL DOTPLOT INDBAL RESOLUTION = 150
/OUTPUT ALL MATCHED.
I get only this output when running the syntax above:
Error
fullmatch unused argument(s) (data = data)
Error in xy.coords(x, y, xlabel, ylabel, log) :
'x' is a list, but does not have components 'x' and 'y'
Nearest neighbor matching seems to work fine, however, Fullmatching doesn't
work either (and Gam logit doesn't work also). The program seems to run fine
in the beginning, but it ends abruptly as stated above in the output. Why
does only nearest neighbor matching seem to work, and how can I make use of
optimal matching and GAM-logit? Does it has to do something with new updates
of R?
Hello,
I agree - the error message is not very helpful and this is something
that we need to work on. Sometimes errors propagate downstream the
code, which results in error messages that pop up at later parts of
the code, even though the problem was further up.
I am not quite sure why the optimal matching is not working for you
and it's difficult to pinpoint. I know for a fact that it worked for
us during testing at one point in time (admittedly, we did have
problems with it too at other times).
I am sorry to say that I am a bit stumped as to what is going on with
the program in your case. Once the semester restarts here in the US we
may have time to explore this further.
Dear Felix,
I tried the past two days getting psmatching to operate properly. I also couldn't find the right solution in the discussion topics as I run through it. Maybe you can help me out?
I installed PSMATCHING3.03 on the fixed IBM SPSS 21.0.0.2, R version 2.14.2, and R essentials for spss 21 on windows xp system 32 bit.
If I try Optimal Matching via the following syntax:
PSMATCHING3
/VARS
ID = ID
TREAT = Controlled_vs_Uncontrolled
COVS = weight length followupDuration
/MATCHIT
MATCH=OPTIMAL
EST =LOGIT
DISCARD = NONE
MORDER = LARGEST
RATIO = 1
CALIPER = 0
/PLOT HISTPLOT JITTERPLOT HISTBAL DOTPLOT INDBAL RESOLUTION = 150
/OUTPUT ALL MATCHED.
I get only this output when running the syntax above:
Error
fullmatch unused argument(s) (data = data)
Error in xy.coords(x, y, xlabel, ylabel, log) :
'x' is a list, but does not have components 'x' and 'y'
Nearest neighbor matching seems to work fine, however, Fullmatching doesn't work either (and Gam logit doesn't work also). The program seems to run fine in the beginning, but it ends abruptly as stated above in the output. Why does only nearest neighbor matching seem to work, and how can I make use of optimal matching and GAM-logit? Does it has to do something with new updates of R?
Thanks in advance for your answer.
This seems to be a bug that we need to look into. I will see what I
can do, we might have to change our code and upload a new version.
On Tue, Jul 8, 2014 at 10:05 AM, Hessel Peters-Sengers
hesselps@users.sf.net wrote:
Hello,
After consulting with my grad student who helped programming, we
believe that the error that optimal / full matching are not working in
your case, might be that the caliper is set to 0. The error message is
indeed not very clear, but if you can change that value and check if
it still not works, let us know. If the caliper is set to 0, only
identical matches may be allowed, which will often result in no
matches at all.
As to whether GAM is not working, I am not sure and we will look into
this. It could also be related to the caliper and less to GAM.
Hope that helps for now,
Felix
On Tue, Jul 8, 2014 at 10:05 AM, Hessel Peters-Sengers
hesselps@users.sf.net wrote:
Dear Felix,
Thanks for your answer. I tried it (set the caliper at different values, also deleted it from the syntax), but I still get the same error:
Error
fullmatch unused argument(s) (data = data)
Error in xy.coords(x, y, xlabel, ylabel, log) :
'x' is a list, but does not have components 'x' and 'y'
I also tried this simplified syntax:
DATASET ACTIVATE DataSet2.
PSMATCHING3
/VARS
ID = ID
TREAT = Controlled_vs_Uncontrolled
COVS = gewicht
/MATCHIT
MATCH=OPTIMAL
EST =LOGIT
DISCARD = NONE
RATIO = 1
CALIPER = .4
/PLOT RESOLUTION = 96
/OUTPUT ALL.
Then I get this error:
Error
fullmatch unused argument(s) (data = data)
Error in data.frame(dict, psspec, weightspec, stringsAsFactors = F) :
arguments imply differing number of rows: 5, 4
It seems the error term gives us not helpful information. Maybe you know how to fix this?
Thanks!
Hello,
I agree - the error message is not very helpful and this is something
that we need to work on. Sometimes errors propagate downstream the
code, which results in error messages that pop up at later parts of
the code, even though the problem was further up.
I am not quite sure why the optimal matching is not working for you
and it's difficult to pinpoint. I know for a fact that it worked for
us during testing at one point in time (admittedly, we did have
problems with it too at other times).
I am sorry to say that I am a bit stumped as to what is going on with
the program in your case. Once the semester restarts here in the US we
may have time to explore this further.
Best,
Felix
On Mon, Jul 14, 2014 at 3:46 AM, Hessel Peters-Sengers
hesselps@users.sf.net wrote: