Menu

General questions

2012-02-27
2013-03-21
  • Yue

    Yue - 2012-09-21

    Hello Felix,

    I just download the psmatching_2 package. however, no files there are .spd files. when I open spss and try to install customer dialog, it can't recongize any file in the package. Do I miss any steps? Thanks

     
  • Felix Thoemmes

    Felix Thoemmes - 2012-09-21

    Hi,
    this is a known problem that occurs because some browsers seem to detect that .spd files are the same as .zip files and rename the downloaded file. To fix this, simply rename the downloaded .zip file into a file with the extension .spd, and everything should work fine.

     
  • Melissa Marselle

    SPSS is not producing propensity scores. R, R plug-in and Felix's plug-in are all installed and working. However, when I run the PS Matching in SPSS, I get one of two error messages, after the 'summary of unbalanced covariates'. The error messages occur in two independent outputs (basically I re-run PS Matching after the first error message to get the second).

    Error message 1:
    Error in OK() :
    (converted from warning) No warnings in estimation or matching procedure
    Error: IBM SPSS Statistics error: The error code is '70' With message 'The label length is too long.'

    Error message 2:
    Error in OK() :
    (converted from warning) No warnings in estimation or matching procedure
    Error: IBM SPSS Statistics error: The error code is '89' With message 'Invalid function call. Function can only be called during dataset creation or within a User Procedure.'

    Any suggestions?

     
    • Felix Thoemmes

      Felix Thoemmes - 2013-01-09

      Hi Melissa,
      unusual SPSS errors that you are getting.
      Not sure if I can pinpoint the problem, but are there any variables or
      value labels that are exceptionally long in your SPSS dataset?
      I would try the following approach that I use to troubleshoot.
      Take a very small test dataset (e.g., just ID, treatment, 2
      covariates) and then see if you get errors with such a minimal
      example. (I have attached a file to your private email).
      If this also yields an error, then there might be a problem with the
      installation of R, the SPSS R essentials Plug-in or an unknown bug in
      my Custom Dialog.
      If this works, there is some issue with your dataset that causes the
      problem. I would be interested to hear if and how you might be able to
      resolve this, so that we can fix any potential bugs in future versions
      of the SPSS custom dialog.

      Best,
      FT

      On Wed, Jan 9, 2013 at 12:06 PM, Melissa Marselle meeza@users.sf.net wrote:

      SPSS is not producing propensity scores. R, R plug-in and Felix's plug-in
      are all installed and working. However, when I run the PS Matching in SPSS,
      I get one of two error messages, after the 'summary of unbalanced
      covariates'. The error messages occur in two independent outputs (basically
      I re-run PS Matching after the first error message to get the second).

      Error message 1:
      Error in OK() :
      (converted from warning) No warnings in estimation or matching procedure
      Error: IBM SPSS Statistics error: The error code is '70' With message 'The
      label length is too long.'

      Error message 2:
      Error in OK() :
      (converted from warning) No warnings in estimation or matching procedure
      Error: IBM SPSS Statistics error: The error code is '89' With message
      'Invalid function call. Function can only be called during dataset creation
      or within a User Procedure.'

      Any suggestions?


      General questions


      Sent from sourceforge.net because you indicated interest in
      https://sourceforge.net/p/psmspss/discussion/general/

      To unsubscribe from further messages, please visit
      https://sourceforge.net/auth/prefs/

       
  • Melissa Marselle

    Dear Felix,

    Thank you for your very quick reply! I really do appreciate it.

    Thank you too for your suggestion. Running the PS Matching programme on a smaller test dataset worked! Encouraged by this, I trimmed my dataset to include ID, treatment variable and all covariates, and deleted all variable labels. PS Matching worked; no SPSS errors!
    I am glad I do not have to learn R!!!! Thank you!

     
  • Vera Feygina

    Vera Feygina - 2013-03-18

    Dear Felix,

    First of all, thanks a lot for this really helpful tool!
    It helped me so much before, but now I faced with some incomprehensible problems. In fact, SPSS doesn't produce any output and doesn't announce any errors. After running the plug-in all I've got is just (even in case of your example file):

    SET SEED = 1234.
    SET PRINTBACK=NONE.

    Sometimes I also receive a warning massage:

    Warning # 6894. Command name: BEGIN PROGRAM
    The external program exit unexpectedly and lost its content, a new exteranl
    program will startup to execute the rest of job.

    I am using SPSS on MacOSX, so I can assume that the problem is somehow about the graphical device -- startx (XQuartz), which opens with the start of the script but doesn't respond.

    Best regards,
    Vera Feygina

     

    Last edit: Vera Feygina 2013-03-18
    • Felix Thoemmes

      Felix Thoemmes - 2013-03-19

      Dear Vera,
      did it work before? I am PC user so some of the errors that you are
      reporting are unfamiliar, especially the Warning # 6894. I have not seen
      this before.
      I would attempt to make sure that the little example that we give in the
      readme.txt is working properly - that way you can test that the SPSS
      R-plugin is working correctly. Then download our minimaltest.sav and try to
      do PSM on this dataset to rule out problems with your data.

      Hope that helps,
      Felix

      P.S. I also asked my colleague Wang Liao, who is a mac user, if he has any
      insights.

      On Mon, Mar 18, 2013 at 7:30 AM, Vera Feygina confidanza@users.sf.netwrote:

      Dear Felix,

      First of all, thanks a lot for this really helpful tool!
      It helped me so much before, but now I faced with some incomprehensible
      problems. In fact, SPSS doesn't produce any output and doesn't announce any
      errors. After running the plug-in all I've got is just (even in case of
      your example file):

      SET SEED = 1234.
      SET PRINTBACK=NONE.
      Sometimes I also receive a warning massage:
      Warning # 6894. Command name: BEGIN PROGRAM
      The external program exit unexpectedly and lost its content, a new
      exteranl
      program will startup to execute the rest of job.

      I am using SPSS on MacOSX, so I can assume that the problem is somehow
      about the graphical device -- startx (XQuartz), which opens with the start
      of the script but doesn't respond.

      General questionshttps://sourceforge.net/p/psmspss/discussion/general/thread/22de6394/?limit=25#622d

      Sent from sourceforge.net because you indicated interest in
      https://sourceforge.net/p/psmspss/discussion/general/

      To unsubscribe from further messages, please visit
      https://sourceforge.net/auth/prefs/

       
  • Vera Feygina

    Vera Feygina - 2013-03-20

    Example from read.me works correctly but PSM on minutest.sav database still produces nothing.

    Felix, thanks a lot for your concern! I will highly appreciate if you or your colleague have any suggestions.

     
    • Felix Thoemmes

      Felix Thoemmes - 2013-03-21

      Hello Vera,
      unfortunately we don't have a solution for your error right now. Given
      that the example from the readme works, but the minimaltest.sav does
      not, it hints at that the PSM program was not correctly installed. Can
      you try to install it again (using an admin account - if such a thing
      exists on a mac), and alternatively also install the old version
      psmatching2 (available under archive).

      Best,
      Felix

      On Wed, Mar 20, 2013 at 5:10 PM, Vera Feygina confidanza@users.sf.net wrote:

      Example from read.me works correctly but PSM on minutest.sav database still
      produces nothing.

      Felix, thanks a lot for your concern! I will highly appreciate if you or
      your colleague have any suggestions.

      General questions


      Sent from sourceforge.net because you indicated interest in
      https://sourceforge.net/p/psmspss/discussion/general/

      To unsubscribe from further messages, please visit
      https://sourceforge.net/auth/prefs/

       

Log in to post a comment.