Missing-Value functions (COMPUTE command)

MISSING VALUE V1 V2 V3 (0).
COMPUTE AllValid=V1 + V2 + V3.
COMPUTE UM=VALUE(V1) + VALUE(V2) + VALUE(V3).
COMPUTE SM=SYSMIS(V1) + SYSMIS(V2) + SYSMIS(V3).
COMPUTE M=MISSING(V1) + MISSING(V2) + MISSING(V3).
  • The MISSING VALUE command declares the value 0 as missing for V1, V2, and V3.
  • AllValid is the sum of three variables only for cases with valid values for all three variables. AllValid is assigned the system-missing value for a case if any variable in the assignment expression has a system- or user-missing value.
  • The VALUE function overrides user-missing value status. Thus, UM is the sum of V1, V2, and V3 for each case, including cases with the value 0 (the user-missing value) for any of the three variables. Cases with the system-missing value for V1, V2, and V3 are system-missing.
  • The SYSMIS function on the third COMPUTE returns the value 1 if the variable is system-missing. Thus, SM ranges from 0 to 3 for each case, depending on whether the variables V1, V2, and V3 are system-missing for that case.
  • The MISSING function on the fourth COMPUTE returns the value 1 if the variable named is system- or user-missing. Thus, M ranges from 0 to 3 for each case, depending on whether the variables V1, V2, and V3 are user- or system-missing for that case.
  • Alternatively, you could use the COUNT command to create the variables SM and M.
    * Test for listwise deletion of missing values.
     
    DATA LIST /V1 TO V6 1-6.
    BEGIN DATA
    213 56
    123457
    123457
    9234 6
    END DATA.
    MISSING VALUES V1 TO V6(6,9).
     
    COMPUTE NotValid=NMISS(V1 TO V6).
    FREQUENCIES VAR=NotValid.
  • COMPUTE determines the number of missing values for each case. For each case without missing values, the value of NotValid is 0. For each case with one missing value, the value of NotValid is 1, and so on. Both system- and user-missing values are counted.
  • FREQUENCIES generates a frequency table for NotValid. The table gives a count of how many cases have all valid values, how many cases have one missing value, how many cases have two missing values, and so on, for variables V1 to V6. This table can be used to determine how many cases would be dropped in an analysis that uses listwise deletion of missing values. For other ways to check listwise deletion, see the examples for the ELSE command (in the DO IF command) and those for the IF command.

See the topic Missing value functions for more information.