P-STAT 2.23 revision 15

P-STAT Version 2.23 revision 15 is latest version available for PC/Windows, Linux and Solaris. This release has no new features but it does contain important changes in the processing of the PSKEY which will make it easier for users to upgrade to version 3.01.

• Revision 14 fixed some minor problems. TABFILE.IN has several enhancements and a bug introduced in revision 13 has been fixed.
  
  
• SURVEY has an enhancement to the CONTENTS capability and a fix for a problem that caused a PPGIFT error message.
  
  
• A minor formatting problem in SPSS.OUT has been fixed.
  

See the news for complete details.

Revision 13 contained the following enhancements.

1. TEXTFILE.IN/TABFILE.IN has been changed. Now the default behaviour is to "honor strings". This is particularly useful when using csv (comma separated variable) files as input. Input values, such as the comma, when found in a string, are now treated as ordinary characters without special meaning. The IGNORE.STRINGS identifier can be used to turn this feature off.
   
   
2. SAMPLE: A SEEDS identifier has been added, which allows the initializing values (three integers from 1 to 30,000) to be specified. This overrides the normal random setting based on date and time.
   
   The final report of a SAMPLE command now shows the initial settings that were used in that run. If a subsequent run is done using those settings as seeds, the same sequence will occur.

Tricks of the trade

1. SAMPLE produces an output file that is a random sample of the input file. This, by itself, makes a useful command. The feature that makes it unique is the ability to specify specific variables in the input file which are to be matched in the output file.
   
   If these "BY" variables are used, each such subgroup will be sampled to the same extent as the file as a whole. There is a limit of 15 BY variables but no limit on the number of subgroups that they define.
   
   
2. EXPAND is the type of function that justifies our belief that P-STAT does indeed have "Simple Solutions for Complex Problems". EXPAND is one of the PPL funtions supported by the menus available in PC/windows.
   
   EXPAND can be used to convert a single variable into a new set of dummy variables. See this example for details.
   
   EXPAND can also be used to project multiple input variables into the set of new variables. In this usage there are several options for the output variables.