SPSS Statistics

SPSS Statistics

Your hub for statistical analysis, data management, and data documentation. Connect, learn, and share with your peers! 

 View Only

  • 1.  Raking gives weights of 1.0000

    Posted Thu March 30, 2023 01:12 PM

    I've used the SPSSINC RAKE procedure for a number of years and came across a new situation today.

    I have a dataset of 500 records.
    When I rake with 1 dimension with two separate variables, I get the results I expect.

    However, when I rake with 2 dimensions with those same two variables, I get weights of 1.0000.




    Any ideas of what might be happening?  Is this a fluke of inverting a matrix or some other mathematical quirk?
    Has anyone else seen this before?
    What do I need to do to "fix" this - assuming that these raked weights are wrong.

    If it helps, here's the code I used:

    weight by wgtEqProb_Hisp_N.
    SPSSINC RAKE 
      DIM1=Area2SexAge5
             111   2999
             112   7496
             113   3911
             114   461
             115   226
             121   2480
             122   6489
             123   2878
             124   253
             125   180
             211   2362
             212   7716
             213   4147
             214   517
             215   177
             221   2629
             222   6671
             223   3921
             224   791
             225   266
      DIM2=EducLevelCategory
              111   696
             112   1372
             113   2329
             121   413
             122   1289
             123   4371
             211   11351
             212   6531
             213   11932
             221   2302
             222   3247
             223   10737
      FINALWEIGHT=raked_weight_Area_Sex_Age_Educ_Hisp
      POPTOTAL=500
      /OPTIONS ITERATIONS=1000 CONVERGENCE=.0001 DELTA=.5.

    Thanks!



    ------------------------------
    Urban Landreman
    ------------------------------



  • 2.  RE: Raking gives weights of 1.0000

    Posted Thu March 30, 2023 01:28 PM
    The test would be if you tabulate the variables with the new weight, do you get the counts you expected?

    If not, can you send me the sav file to investigate (jkpeck@gmail.com)?

    --



    Original Message


  • 3.  RE: Raking gives weights of 1.0000

    Posted Thu March 30, 2023 03:23 PM

    I forgot to mention that I am using SPSS Statistics Release 26.0.0.0, 64-bit edition.



    ------------------------------
    Urban Landreman
    ------------------------------

    Original Message


Related Content