Yes, I recently noticed this issue too and first thought that it was a
compatability problem.
When I encountered this problem of not being able to install the 'cem'
package, I also verified on CRAN that it is up-to-date (which it was).
And yet, even from within R (not via the SPSS R-Plugin), I was not
able to install 'cem'.
This manual workaround helps:
Just so you know, after following your instructions, I first got this:
ERROR: dependency 'randomForest' is not available for package 'cem'
After manually installing randomForest, installing cem worked.
I then ran PS Matching in SPSS, during which R automatically installed three more packages (SparseM, xtable, and RItools).
Thanks again!!!
Heiko
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi Felix,
I have the same problem as Heiko. I am trying to get a PS matching integrated into SPSS but I seem to have problems to get the R and R essentials to work. When I try to install cem it says that it is no available for 2.14.2. I have SPSS version 21. R seems not not to understand the command
install.packages("C:/Downloads/cem_1.1.5.tar.gz", repos = NULL, type = "source")
Sorry to ask, but do you need to type repos=NULL, type ="source") precisely as it says. It seems that I dont understand this part. I have been tryng for 12 days without succeding. Please help me!
Kind regards
Ashkan
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Dear Ashkan,
sorry to hear that you had a hard time installing the program.
Yes, you need to include repos and type statements, as they basically
tell R not to download "cem" from the CRAN server, but from your local
harddrive.
You also need to download an older version (e.g., 1.1.2) of "cem" here: http://gking.harvard.edu/cem
We are almost finished with our new version of PSMATCHING that will
once and for all fix the "cem" installation problem that many have
encountered.
Hi Felix,
I have the same problem as Heiko. I am trying to get a PS matching
integrated into SPSS but I seem to have problems to get the R and R
essentials to work. When I try to install cem it says that it is no
available for 2.14.2. I have SPSS version 21. R seems not not to understand
the command
install.packages("C:/Downloads/cem_1.1.5.tar.gz", repos = NULL, type =
"source")
Sorry to ask, but do you need to type repos=NULL, type ="source") precisely
as it says. It seems that I dont understand this part. I have been tryng for
12 days without succeding. Please help me!
Kind regards
Ashkan
Hi Felix,
Thank you for your quick response! I have tried with cem 1.1.2 as well. I get the feeling that I am typing this specific part incorrectly.
install.packages("C:/Downloads/cem_1.1.2.tar_0.gz", repos = NULL, type =
"source")
I change it off course to the place where I download it from. Should there be a space after "," and after and before "=". I have tried all possible combinations.
Appreciate the help!
Best,
Ashkan
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi Ashkan,
not sure why this is not working. The spacing should not matter.
Maybe this post will be helpful in explaining how to install a package
that you downloaded.
Hi Felix,
Thank you for your quick response! I have tried with cem 1.1.2 as well. I
get the feeling that I am typing this specific part incorrectly.
install.packages("C:/Downloads/cem_1.1.2.tar_0.gz", repos = NULL, type =
"source")
I change it off course to the place where I download it from. Should there
be a space after "," and after and before "=". I have tried all possible
combinations.
Appreciate the help!
I have figured out this part. I am able to install all the packages that are recommended including randomForest, "combinat" and the cem 1.1.2 and 1.1.4 (tried both) without success. It seems that R 2.14.2 and the essentials for R 21 are not compatible. When I type (SPSS syntax)
x <- "R plug-in is working properly"
x
END PROGRAM.
I get this response from SPSS no matter which of the versions I use.
"The first word in the line is not recognised as an SPSS statistics command. "
I think that this is the reason why PS matching is not working. I have also tried the latest R version with the same essentials for R 21. I have the version 21 of SPSS. Sorry to write again, but I am working against a deadline. Would appreciate it for the world if you could help me with this one.
Best,
Ashkan
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi Ashkan,
based on your description, this is definitely a problem with the SPSS
R-Essentials plugin.
Have you checked the SPSS documentation on how to install the R-plugin?
ftp://public.dhe.ibm.com/software/analytics/spss/documentation/statistics/21.0/en/rplugin/InstallationDocuments/Windows/Essentials_for_R_Installation_Instructions.pdf
I have figured out this part. I am able to install all the packages that are
recommended including randomForest, "combinat" and the cem 1.1.2 and 1.1.4
(tried both) without success. It seems that R 2.14.2 and the essentials for
R 21 are not compatible. When I type (SPSS syntax)
x <- "R plug-in is working properly"
x
END PROGRAM.
I get this response from SPSS no matter which of the versions I use.
"The first word in the line is not recognised as an SPSS statistics command.
"
I think that this is the reason why PS matching is not working. I have also
tried the latest R version with the same essentials for R 21. I have the
version 21 of SPSS. Sorry to write again, but I am working against a
deadline. Would appreciate it for the world if you could help me with this
one.
Hi Felix,
I have read that one. I am using windows 8 and I run the R-plug as an administrator. Simply wont work! Get the same response in the syntax. All the version should match. Other leads?
Best,
Ashkan
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Never tested windows 8 before - this could be a potential issue, even
though it really shouldn't.
A couple of suggestions: check the SPSS developer forum and post your
question there. Since this is a problem with the communication between
the R-plugin and SPSS, they might be able to help you. Jon Peck is
active on these forums and he is very knowledgeable. https://www.ibm.com/developerworks/community/forums/html/category?id=33333333-0000-0000-0000-000000000302
Second solution would be to switch completely to R and use the
MatchIt() package.
Finally, John Painter (UNC) has an SPSS macro (not a dialog or
extension) that performs simple propensity score matching.
Hi Felix,
I have read that one. I am using windows 8 and I run the R-plug as an
administrator. Simply wont work! Get the same response in the syntax. All
the version should match. Other leads?
Hi Felix,
I managed to get R and plug-in work by marking the whole lines to the left in the syntax (I am sorry, but I am beginner). But the problem remains with PS matching. I have installed cem_1.1.2, randomForest, and all the packages that the output suggests. This is the response in the output window when I try to do some calculation in PS matching:
or help.matchit() for help with commands supported by MatchIt.
Package SparseM (0.97) loaded.
To cite, see citation("SparseM")
Loading required package: nlme
Loading required package: lattice
Loading required package: randomForest
Error: package 'randomForest' could not be loaded
In addition: Warning message:
In library(pkg, character.only = TRUE, logical.return = TRUE, lib.loc = lib.loc) :
there is no package called 'randomForest'
Error: object 'pt3' not found
Loading required package: randomForest
Error: package 'randomForest' could not be loaded
In addition: Warning message:
In library(pkg, character.only = TRUE, logical.return = TRUE, lib.loc = lib.loc) :
there is no package called 'randomForest'
Warning message:
In OK() : No warnings in estimation or matching procedure
Error: unexpected input in:
I have even typed install.packages("MatchIt") which installs correctly I guess. Any other clues?
Best,
Ashkan
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hello Ashkan,
it's still not easy to pinpoint, but it seems that the problem is with
the "random forest" package.
Two suggestions: uninstall and reinstall the "random forest" package
directly in R.
Alternatively, contact me off-list and I can send you the pre-release
version of PSMATCHING3 that may work better with your system.
Best,
Felix
Hi Felix,
I managed to get R and plug-in work by marking the whole lines to the left
in the syntax (I am sorry, but I am beginner). But the problem remains with
PS matching. I have installed cem_1.1.2, randomForest, and all the packages
that the output suggests. This is the response in the output window when I
try to do some calculation in PS matching:
or help.matchit() for help with commands supported by MatchIt.
Package SparseM (0.97) loaded.
To cite, see citation("SparseM")
Loading required package: nlme
Loading required package: lattice
Loading required package: randomForest
Error: package 'randomForest' could not be loaded
In addition: Warning message:
In library(pkg, character.only = TRUE, logical.return = TRUE, lib.loc =
lib.loc) :
there is no package called 'randomForest'
Error: object 'pt3' not found
Loading required package: randomForest
Error: package 'randomForest' could not be loaded
In addition: Warning message:
In library(pkg, character.only = TRUE, logical.return = TRUE, lib.loc =
lib.loc) :
there is no package called 'randomForest'
Warning message:
In OK() : No warnings in estimation or matching procedure
Error: unexpected input in:
I have even typed install.packages("MatchIt") which installs correctly I
guess. Any other clues?
Best,
Ashkan
Hi Felix!
I would really appreciate that. But I have some difficulties to find your email. Please e-amil at ashkan.labaf@gmail.com.
Thanks a lot!
Best,
Ashkan
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi Felix,
when I try to run PS Matching in SPSS 21 (R 2.14.2), I get the following error:
Trying to manually install cem leads to:
Interestingly, cem appears to be up-to-date: http://cran.r-project.org/web/packages/cem/index.html
Maybe I need an older version?
Thanks for all the good work!
Best,
Heiko
Hi Heiko,
thanks for the post.
Yes, I recently noticed this issue too and first thought that it was a
compatability problem.
When I encountered this problem of not being able to install the 'cem'
package, I also verified on CRAN that it is up-to-date (which it was).
And yet, even from within R (not via the SPSS R-Plugin), I was not
able to install 'cem'.
This manual workaround helps:
1.) Download the 'cem' package here
http://gking.harvard.edu/cem
2.) Open R and type
install.packages("combinat")
install.packages("C:/Downloads/cem_1.1.5.tar.gz", repos = NULL, type = "source")
where you need to replace C:/Downloads with the path where you have
downloaded the tar.gz file.
Let me know if that does not work for you.
Best,
Felix
On Wed, Apr 17, 2013 at 4:13 AM, Heiko hbreitsohl@users.sf.net wrote:
Hi Felix,
it appears to work now. Yay!
Just so you know, after following your instructions, I first got this:
ERROR: dependency 'randomForest' is not available for package 'cem'
After manually installing randomForest, installing cem worked.
I then ran PS Matching in SPSS, during which R automatically installed three more packages (SparseM, xtable, and RItools).
Thanks again!!!
Heiko
Hi Felix,
I have the same problem as Heiko. I am trying to get a PS matching integrated into SPSS but I seem to have problems to get the R and R essentials to work. When I try to install cem it says that it is no available for 2.14.2. I have SPSS version 21. R seems not not to understand the command
install.packages("C:/Downloads/cem_1.1.5.tar.gz", repos = NULL, type = "source")
Sorry to ask, but do you need to type repos=NULL, type ="source") precisely as it says. It seems that I dont understand this part. I have been tryng for 12 days without succeding. Please help me!
Kind regards
Ashkan
Dear Ashkan,
sorry to hear that you had a hard time installing the program.
Yes, you need to include repos and type statements, as they basically
tell R not to download "cem" from the CRAN server, but from your local
harddrive.
You also need to download an older version (e.g., 1.1.2) of "cem" here:
http://gking.harvard.edu/cem
We are almost finished with our new version of PSMATCHING that will
once and for all fix the "cem" installation problem that many have
encountered.
Best,
Felix
On Mon, Jun 17, 2013 at 3:07 PM, Ashkan Labaf ash21@users.sf.net wrote:
Hi Felix,
Thank you for your quick response! I have tried with cem 1.1.2 as well. I get the feeling that I am typing this specific part incorrectly.
install.packages("C:/Downloads/cem_1.1.2.tar_0.gz", repos = NULL, type =
"source")
I change it off course to the place where I download it from. Should there be a space after "," and after and before "=". I have tried all possible combinations.
Appreciate the help!
Best,
Ashkan
Hi Ashkan,
not sure why this is not working. The spacing should not matter.
Maybe this post will be helpful in explaining how to install a package
that you downloaded.
http://stackoverflow.com/questions/1474081/how-do-i-install-an-r-package-from-source
Best,
Felix
On Tue, Jun 18, 2013 at 2:23 AM, Ashkan Labaf ash21@users.sf.net wrote:
Hi again Felix,
I have figured out this part. I am able to install all the packages that are recommended including randomForest, "combinat" and the cem 1.1.2 and 1.1.4 (tried both) without success. It seems that R 2.14.2 and the essentials for R 21 are not compatible. When I type (SPSS syntax)
x <- "R plug-in is working properly"
x
END PROGRAM.
I get this response from SPSS no matter which of the versions I use.
"The first word in the line is not recognised as an SPSS statistics command. "
I think that this is the reason why PS matching is not working. I have also tried the latest R version with the same essentials for R 21. I have the version 21 of SPSS. Sorry to write again, but I am working against a deadline. Would appreciate it for the world if you could help me with this one.
Best,
Ashkan
Hi Ashkan,
based on your description, this is definitely a problem with the SPSS
R-Essentials plugin.
Have you checked the SPSS documentation on how to install the R-plugin?
ftp://public.dhe.ibm.com/software/analytics/spss/documentation/statistics/21.0/en/rplugin/InstallationDocuments/Windows/Essentials_for_R_Installation_Instructions.pdf
Best,
Felix
On Wed, Jun 19, 2013 at 2:24 PM, Ashkan Labaf ash21@users.sf.net wrote:
Hi Felix,
I have read that one. I am using windows 8 and I run the R-plug as an administrator. Simply wont work! Get the same response in the syntax. All the version should match. Other leads?
Best,
Ashkan
Never tested windows 8 before - this could be a potential issue, even
though it really shouldn't.
A couple of suggestions: check the SPSS developer forum and post your
question there. Since this is a problem with the communication between
the R-plugin and SPSS, they might be able to help you. Jon Peck is
active on these forums and he is very knowledgeable.
https://www.ibm.com/developerworks/community/forums/html/category?id=33333333-0000-0000-0000-000000000302
Second solution would be to switch completely to R and use the
MatchIt() package.
Finally, John Painter (UNC) has an SPSS macro (not a dialog or
extension) that performs simple propensity score matching.
Hope that helps,
Felix
On Wed, Jun 19, 2013 at 2:53 PM, Ashkan Labaf ash21@users.sf.net wrote:
Hi Felix,
I managed to get R and plug-in work by marking the whole lines to the left in the syntax (I am sorry, but I am beginner). But the problem remains with PS matching. I have installed cem_1.1.2, randomForest, and all the packages that the output suggests. This is the response in the output window when I try to do some calculation in PS matching:
Loading required package: MASS
MatchIt (Version 2.4-20, built: 2011-10-24)
Please refer to http://gking.harvard.edu/matchit for full documentation
or help.matchit() for help with commands supported by MatchIt.
Package SparseM (0.97) loaded.
To cite, see citation("SparseM")
Loading required package: nlme
Loading required package: lattice
Loading required package: randomForest
Error: package 'randomForest' could not be loaded
In addition: Warning message:
In library(pkg, character.only = TRUE, logical.return = TRUE, lib.loc = lib.loc) :
there is no package called 'randomForest'
Error: object 'pt3' not found
Loading required package: randomForest
Error: package 'randomForest' could not be loaded
In addition: Warning message:
In library(pkg, character.only = TRUE, logical.return = TRUE, lib.loc = lib.loc) :
there is no package called 'randomForest'
Warning message:
In OK() : No warnings in estimation or matching procedure
Error: unexpected input in:
I have even typed install.packages("MatchIt") which installs correctly I guess. Any other clues?
Best,
Ashkan
Hello Ashkan,
it's still not easy to pinpoint, but it seems that the problem is with
the "random forest" package.
Two suggestions: uninstall and reinstall the "random forest" package
directly in R.
Alternatively, contact me off-list and I can send you the pre-release
version of PSMATCHING3 that may work better with your system.
Best,
Felix
On Wed, Jun 19, 2013 at 5:53 PM, Ashkan Labaf ash21@users.sf.net wrote:
Hi Felix!
I would really appreciate that. But I have some difficulties to find your email. Please e-amil at ashkan.labaf@gmail.com.
Thanks a lot!
Best,
Ashkan