SPSS Statistics

Robert Lundqvist posted Fri February 07, 2025 10:08 AM

It seems as if your making a table for the old variable, not the new recoded one. The line "frequencies variable_name", shouldn't it be "frequencies response_order_numeric"?

Jon Peck IBM Champion posted Fri February 07, 2025 10:24 AM

Frequencies does not give you custom control over the category order other than by count or alpha order, but Custom Tables lets you specify the category order.  See the  Categories and Totals subdialog

Here is sample syntax

CTABLES
  /TABLE jobcat
  /CATEGORIES VARIABLES=jobcat [2, 3, 1, OTHERNM] EMPTY=INCLUDE.

Art Jack posted Fri February 07, 2025 10:26 AM

you could also recode the string to a numeric value & place the corresponding value labels on top.  they'll show up in numeric order.

Andrew Munn posted Tue February 11, 2025 03:11 PM

Thanks for your help, everyone. It turns out the SPSS license I have through my job doesn't let me use custom tables (that costs extra, I guess), but recoding the string variable to a new numeric variable and adding in the original variable's categories as value labels worked. For anyone who has this problem in the future, this is how I ended up doing it:

missing values variable_name ('').

recode variable_name

    ("Extremely confident" = 1)

    ("Confident" = 2)

    ("Neither confident nor unconfident" = 3)

    ("Unconfident" = 4)

    ("Extremely unconfident" = 5)

    (convert) into variable_name_numeric.

value labels variable_name_numeric    

    1 "Extremely confident"

    2 "Confident"

    3 "Neither confident nor unconfident"

    4 "Unconfident"

    5 "Extremely unconfident".

sort cases by variable_name_numeric(A).

frequencies variable_name_numeric.

execute.

Jon Peck IBM Champion posted Tue February 11, 2025 04:02 PM

CTABLES is part of the Standard Edition of SPSS.  There is a Base version that is cheaper, but it might be worth considering moving up to the Standard Edition if you use Statostics a lot.