Home‎ > ‎Statistics‎ > ‎SPSS‎ > ‎My SPSS Page‎ > ‎

rhoCI

An SPSS Macro to Compute Confidence Intervals for Pearson's Correlation

Ray Koopman and I wrote an article called An SPSS Macro to Compute Confidence Intervals for Pearson's Correlation.  We wrote it because the SPSS CORRELATIONS procedure does not provide confidence intervals (CIs) for rho.  The article has now been published in The Quantitative Methods for Psychology, and can be downloaded free of charge here.  The macro definition file and a syntax file showing examples of its use are available via the links provided below.

The macro definition file and demonstration files have been uploaded as text files (with .txt as the file extension).  Left-clicking on the links for those files will cause them to open in your text editor (or possibly in your web browser).  This was done to facilitate viewing of the files by folks who do not have SPSS installed.  To download those files as SPSS syntax files, right-click on the links and select Save As (or Save Link As).  Then add the extension .SPS to the file name, and enclose the entire file name in quotes -- e.g., "rhoCI.SPS".
  • The article. [pdf]
  • The macro definition. (Right-click and select Save Link As to save as a syntax file.)
  • Examples of how to use the macro. (Right-click and select Save Link As to save as a syntax file.) 
  • Output from the examples. [pdf]

NOTE:  The macro definition and example files are actually stored as text files, with .txt as the file extension.  To save them as SPSS syntax files, include .SPS as the file extension in the file-save dialog, and enclose the entire file name in quotes.  For example, when saving the macro definition file, change rhoCI to "rhoCI.SPS", including the quotes.  If you leave it as rhoCI, the downloaded file will be a text file called rhoCI.txt.

Also, Jon Peck alerted me to the fact that the !rhoCI macro did not work when he handed it a variable list consisting of X and Y.  It turned out that the problem was due to the use of X and Y as variable names in the macro code.  The now modified macro definition uses @X@ and @Y@ rather than X and Y as variable names.  So provided you don't have variables of your own named @X@ or @Y@, everything should work just fine.  Thanks for spotting that bug, Jon.



Last updated: 14-Jul-2016
Comments