PM58893: MERSENNE TWISTER GIVES INCORRECT SEED NUMBER ON CROSSTABS FOOTNOTE FOR MONTE CARLO EXACT TEST ON WINDOWS 7 64 BIT

APAR status

• Closed as program error.

Error description

• You work with IBM SPSS Statistics 20 on a Windows 7 64 bit
  operating system.
  You create a crosstab with Chi Square test and Monte Carlo exact
  test. For the random number generation you have activated
  Mersenne Twister with:
  
  SET RNG=MT MTINDEX=2000000.
  
  However, on 64 bit OS the footnote you get on the Chi Square
  table is incorrect for the starting seed number. For example you
  get a value of 2673618506938605000:
  'b. Based on 10000 sampled tables with starting seed
  2673618506938605000'
  
  This is not an exact number because a seed number can be up to
  2000000000. This does not happen on a 32 bit machine.
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED:                                              *
  * Users working with Mersenne Twister for Exact Tests on 64    *
  * bit OS                                                       *
  ****************************************************************
  * PROBLEM DESCRIPTION:                                         *
  * You work with IBM SPSS Statistics 20 on a Windows 7 64 bit   *
  * operating system.                                            *
  * You create a crosstab with Chi Square test and Monte Carlo   *
  * exact                                                        *
  * test. For the random number generation you have activated    *
  * Mersenne Twister with:                                       *
  *                                                              *
  * SET RNG=MT MTINDEX=2000000.                                  *
  *                                                              *
  * However, on 64 bit OS the footnote you get on the Chi Square *
  * table is incorrect for the starting seed number. For example *
  * you                                                          *
  * get a value of 2673618506938605000:                          *
  * 'b. Based on 10000 sampled tables with starting seed         *
  * 2673618506938605000'                                         *
  *                                                              *
  * This is not an exact number because a seed number can be up  *
  * to                                                           *
  * 2000000000. This does not happen on a 32 bit machine.        *
  ****************************************************************
  * RECOMMENDATION:                                              *
  * This issue has been addressed in IBM SPSS Statistics Fixpack *
  * 2 (20.0.0.2)                                                 *
  ****************************************************************
  

Problem conclusion

• This issue has been addressed in IBM SPSS Statistics Fixpack 2
  (20.0.0.2)
  

Temporary fix

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

Fix information

• Fixed component name

  SPSS STATISTICS

• Fixed component ID

  5725A54ST

Applicable component levels

• RK00 PSN

     UP

• RK00 PSY

     UP