Missing values can be either system missing values or defined by the user. Basically there are two different types of system missing values:
+ skipped questions (if a certain question was asked only of men, then all women will have a missing value for that variable)
+ dropout questions (if a certain question was asked to all respondents but someone allowed to answer, did not respond).
You, as a user can also define missing values. If you do not want the answer 'Don't know' to be included in your tables and in calculations, you can declare that particular category as missing for the concerned variable.
Missing values are values that receive special treatment when creating tables, in the calculation of statistics, data transformation and the selection of cases. Missing values by default are not included in tables, and are also not included in the calculation of statistics such as the mean or the percentage (it is possible to include missing values in tables, using the table options menu).
Declaring missing values for a variable
You can declare one or more values missing for one variable by double-clicking over the variable name in Variable view's list.
In the edit box labelled "Missings" You can add a comma separated list of values that will be treated as missing for this variable.
NEW: You can also declare missing values for text variables.
You can manage missing values by selecting the variables that you want to manage, and then click "Manage missings".
Next you can choose to:
- Remove all missing value definitions from the selected variables
- Remove missing values from the selected variables using a mask, for example if you have missing values defined as 997, 998 you can remove them using a mask "99?" or "99*"
- Add missing values by specifying a comma separated list of values or masks. Specifying "9991,9992, 999?" will add values 9991 and 9992 and any other 4 digit value starting with '999' for which a label is defined.