I have a data set with information on BMI which is derived either from self-reported (variable 1) or measured weight (variable 2). Each subject has information on BMI in one of these variables, not both. How do I combine these into variable 3, 'any BMI'? I do not use syntax, but commands.
Of course it is possible to calculate an 'any BMI' variable by using commands (transform -> compute variable), but to my opinion it meight be much easier to use the syntax. For example:
NUMERIC bmi_any (F2.1).
COMPUTE bmi_any = bmi_self.
IF NOT MISSING (bmi_meas) bmi_any = bmi_meas.
EXECUTE.
bmi_any is the new variable, bmi_self the self-reported and bmi_meas the measured value. This way combines both values in bmi_any variable. In this way, the new variable is calculated hierarchically. This means that if a case has values in both variants, the measured value is used.
Of course it is possible to calculate an 'any BMI' variable by using commands (transform -> compute variable), but to my opinion it meight be much easier to use the syntax. For example:
NUMERIC bmi_any (F2.1).
COMPUTE bmi_any = bmi_self.
IF NOT MISSING (bmi_meas) bmi_any = bmi_meas.
EXECUTE.
bmi_any is the new variable, bmi_self the self-reported and bmi_meas the measured value. This way combines both values in bmi_any variable. In this way, the new variable is calculated hierarchically. This means that if a case has values in both variants, the measured value is used.
Dear Heidi,
If you wont to do that in a more easy or without spss syntax or commands, you can use the concatenate function in Excel. It's very easy because you have to put in one new cell CONCATENATE(A1;B1) , where A1 is the first column ore the first variable (in row 1 ore case 1) a B1 is the second column ore the second variable (in row 1 or case 1). With this, you create a new variable that contains the first BMI measures or the second BMI measures, depending on case responses. Then, you can easily add this new variable in spss, just cutting and pasting this.
- Badges
- Science topic